diff --git "a/model.safetensors.index.json" "b/model.safetensors.index.json" new file mode 100644--- /dev/null +++ "b/model.safetensors.index.json" @@ -0,0 +1,2490 @@ +{ + "metadata": { + "total_size": 17680980288 + }, + "weight_map": { + "model.embed_tokens.weight": "model-00001-of-00004.safetensors", + "model.final_norm.weight": "model-00004-of-00004.safetensors", + "model.layers.0.input_norm.weight": "model-00001-of-00004.safetensors", + "model.layers.0.res_scale.hidden_states_bias": "model-00001-of-00004.safetensors", + "model.layers.0.res_scale.hidden_states_scale": "model-00001-of-00004.safetensors", + "model.layers.0.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.0.self_attn.qkv.conv_qk.0.bias": "model-00001-of-00004.safetensors", + "model.layers.0.self_attn.qkv.conv_qk.0.weight": "model-00001-of-00004.safetensors", + "model.layers.0.self_attn.qkv.conv_qk.1.bias": "model-00001-of-00004.safetensors", + "model.layers.0.self_attn.qkv.conv_qk.1.weight": "model-00001-of-00004.safetensors", + "model.layers.0.self_attn.qkv.linear_k.weight": "model-00001-of-00004.safetensors", + "model.layers.0.self_attn.qkv.linear_q.weight": "model-00001-of-00004.safetensors", + "model.layers.0.self_attn.qkv.temp": "model-00001-of-00004.safetensors", + "model.layers.0.self_attn.qkv.val_proj1.weight": "model-00001-of-00004.safetensors", + "model.layers.0.self_attn.qkv.val_proj2.weight": "model-00001-of-00004.safetensors", + "model.layers.1.input_norm.weight": "model-00001-of-00004.safetensors", + "model.layers.1.res_scale.hidden_states_bias": "model-00001-of-00004.safetensors", + "model.layers.1.res_scale.hidden_states_scale": "model-00001-of-00004.safetensors", + "model.layers.1.res_scale.residual_bias": "model-00001-of-00004.safetensors", + "model.layers.1.res_scale.residual_scale": "model-00001-of-00004.safetensors", + "model.layers.1.zaya_block.experts.local_experts.0.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.1.zaya_block.experts.local_experts.0.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.1.zaya_block.experts.local_experts.1.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.1.zaya_block.experts.local_experts.1.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.1.zaya_block.experts.local_experts.10.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.1.zaya_block.experts.local_experts.10.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.1.zaya_block.experts.local_experts.11.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.1.zaya_block.experts.local_experts.11.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.1.zaya_block.experts.local_experts.12.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.1.zaya_block.experts.local_experts.12.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.1.zaya_block.experts.local_experts.13.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.1.zaya_block.experts.local_experts.13.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.1.zaya_block.experts.local_experts.14.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.1.zaya_block.experts.local_experts.14.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.1.zaya_block.experts.local_experts.15.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.1.zaya_block.experts.local_experts.15.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.1.zaya_block.experts.local_experts.2.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.1.zaya_block.experts.local_experts.2.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.1.zaya_block.experts.local_experts.3.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.1.zaya_block.experts.local_experts.3.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.1.zaya_block.experts.local_experts.4.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.1.zaya_block.experts.local_experts.4.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.1.zaya_block.experts.local_experts.5.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.1.zaya_block.experts.local_experts.5.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.1.zaya_block.experts.local_experts.6.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.1.zaya_block.experts.local_experts.6.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.1.zaya_block.experts.local_experts.7.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.1.zaya_block.experts.local_experts.7.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.1.zaya_block.experts.local_experts.8.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.1.zaya_block.experts.local_experts.8.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.1.zaya_block.experts.local_experts.9.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.1.zaya_block.experts.local_experts.9.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.1.zaya_block.router.balancing_biases": "model-00001-of-00004.safetensors", + "model.layers.1.zaya_block.router.down_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.1.zaya_block.router.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.1.zaya_block.router.rmsnorm_eda.weight": "model-00001-of-00004.safetensors", + "model.layers.1.zaya_block.router.router_mlp.0.bias": "model-00001-of-00004.safetensors", + "model.layers.1.zaya_block.router.router_mlp.0.weight": "model-00001-of-00004.safetensors", + "model.layers.1.zaya_block.router.router_mlp.2.bias": "model-00001-of-00004.safetensors", + "model.layers.1.zaya_block.router.router_mlp.2.weight": "model-00001-of-00004.safetensors", + "model.layers.1.zaya_block.router.router_mlp.4.weight": "model-00001-of-00004.safetensors", + "model.layers.10.input_norm.weight": "model-00001-of-00004.safetensors", + "model.layers.10.res_scale.hidden_states_bias": "model-00001-of-00004.safetensors", + "model.layers.10.res_scale.hidden_states_scale": "model-00001-of-00004.safetensors", + "model.layers.10.res_scale.residual_bias": "model-00001-of-00004.safetensors", + "model.layers.10.res_scale.residual_scale": "model-00001-of-00004.safetensors", + "model.layers.10.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.10.self_attn.qkv.conv_qk.0.bias": "model-00001-of-00004.safetensors", + "model.layers.10.self_attn.qkv.conv_qk.0.weight": "model-00001-of-00004.safetensors", + "model.layers.10.self_attn.qkv.conv_qk.1.bias": "model-00001-of-00004.safetensors", + "model.layers.10.self_attn.qkv.conv_qk.1.weight": "model-00001-of-00004.safetensors", + "model.layers.10.self_attn.qkv.linear_k.weight": "model-00001-of-00004.safetensors", + "model.layers.10.self_attn.qkv.linear_q.weight": "model-00001-of-00004.safetensors", + "model.layers.10.self_attn.qkv.temp": "model-00001-of-00004.safetensors", + "model.layers.10.self_attn.qkv.val_proj1.weight": "model-00001-of-00004.safetensors", + "model.layers.10.self_attn.qkv.val_proj2.weight": "model-00001-of-00004.safetensors", + "model.layers.11.input_norm.weight": "model-00001-of-00004.safetensors", + "model.layers.11.res_scale.hidden_states_bias": "model-00001-of-00004.safetensors", + "model.layers.11.res_scale.hidden_states_scale": "model-00001-of-00004.safetensors", + "model.layers.11.res_scale.residual_bias": "model-00001-of-00004.safetensors", + "model.layers.11.res_scale.residual_scale": "model-00001-of-00004.safetensors", + "model.layers.11.zaya_block.experts.local_experts.0.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.11.zaya_block.experts.local_experts.0.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.11.zaya_block.experts.local_experts.1.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.11.zaya_block.experts.local_experts.1.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.11.zaya_block.experts.local_experts.10.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.11.zaya_block.experts.local_experts.10.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.11.zaya_block.experts.local_experts.11.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.11.zaya_block.experts.local_experts.11.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.11.zaya_block.experts.local_experts.12.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.11.zaya_block.experts.local_experts.12.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.11.zaya_block.experts.local_experts.13.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.11.zaya_block.experts.local_experts.13.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.11.zaya_block.experts.local_experts.14.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.11.zaya_block.experts.local_experts.14.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.11.zaya_block.experts.local_experts.15.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.11.zaya_block.experts.local_experts.15.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.11.zaya_block.experts.local_experts.2.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.11.zaya_block.experts.local_experts.2.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.11.zaya_block.experts.local_experts.3.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.11.zaya_block.experts.local_experts.3.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.11.zaya_block.experts.local_experts.4.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.11.zaya_block.experts.local_experts.4.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.11.zaya_block.experts.local_experts.5.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.11.zaya_block.experts.local_experts.5.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.11.zaya_block.experts.local_experts.6.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.11.zaya_block.experts.local_experts.6.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.11.zaya_block.experts.local_experts.7.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.11.zaya_block.experts.local_experts.7.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.11.zaya_block.experts.local_experts.8.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.11.zaya_block.experts.local_experts.8.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.11.zaya_block.experts.local_experts.9.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.11.zaya_block.experts.local_experts.9.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.11.zaya_block.router.balancing_biases": "model-00001-of-00004.safetensors", + "model.layers.11.zaya_block.router.down_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.11.zaya_block.router.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.11.zaya_block.router.rmsnorm_eda.weight": "model-00001-of-00004.safetensors", + "model.layers.11.zaya_block.router.router_mlp.0.bias": "model-00001-of-00004.safetensors", + "model.layers.11.zaya_block.router.router_mlp.0.weight": "model-00001-of-00004.safetensors", + "model.layers.11.zaya_block.router.router_mlp.2.bias": "model-00001-of-00004.safetensors", + "model.layers.11.zaya_block.router.router_mlp.2.weight": "model-00001-of-00004.safetensors", + "model.layers.11.zaya_block.router.router_mlp.4.weight": "model-00001-of-00004.safetensors", + "model.layers.11.zaya_block.router.router_states_scale": "model-00001-of-00004.safetensors", + "model.layers.12.input_norm.weight": "model-00001-of-00004.safetensors", + "model.layers.12.res_scale.hidden_states_bias": "model-00001-of-00004.safetensors", + "model.layers.12.res_scale.hidden_states_scale": "model-00001-of-00004.safetensors", + "model.layers.12.res_scale.residual_bias": "model-00001-of-00004.safetensors", + "model.layers.12.res_scale.residual_scale": "model-00001-of-00004.safetensors", + "model.layers.12.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.12.self_attn.qkv.conv_qk.0.bias": "model-00001-of-00004.safetensors", + "model.layers.12.self_attn.qkv.conv_qk.0.weight": "model-00001-of-00004.safetensors", + "model.layers.12.self_attn.qkv.conv_qk.1.bias": "model-00001-of-00004.safetensors", + "model.layers.12.self_attn.qkv.conv_qk.1.weight": "model-00001-of-00004.safetensors", + "model.layers.12.self_attn.qkv.linear_k.weight": "model-00001-of-00004.safetensors", + "model.layers.12.self_attn.qkv.linear_q.weight": "model-00001-of-00004.safetensors", + "model.layers.12.self_attn.qkv.temp": "model-00001-of-00004.safetensors", + "model.layers.12.self_attn.qkv.val_proj1.weight": "model-00001-of-00004.safetensors", + "model.layers.12.self_attn.qkv.val_proj2.weight": "model-00001-of-00004.safetensors", + "model.layers.13.input_norm.weight": "model-00001-of-00004.safetensors", + "model.layers.13.res_scale.hidden_states_bias": "model-00001-of-00004.safetensors", + "model.layers.13.res_scale.hidden_states_scale": "model-00001-of-00004.safetensors", + "model.layers.13.res_scale.residual_bias": "model-00001-of-00004.safetensors", + "model.layers.13.res_scale.residual_scale": "model-00001-of-00004.safetensors", + "model.layers.13.zaya_block.experts.local_experts.0.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.13.zaya_block.experts.local_experts.0.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.13.zaya_block.experts.local_experts.1.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.13.zaya_block.experts.local_experts.1.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.13.zaya_block.experts.local_experts.10.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.13.zaya_block.experts.local_experts.10.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.13.zaya_block.experts.local_experts.11.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.13.zaya_block.experts.local_experts.11.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.13.zaya_block.experts.local_experts.12.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.13.zaya_block.experts.local_experts.12.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.13.zaya_block.experts.local_experts.13.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.13.zaya_block.experts.local_experts.13.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.13.zaya_block.experts.local_experts.14.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.13.zaya_block.experts.local_experts.14.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.13.zaya_block.experts.local_experts.15.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.13.zaya_block.experts.local_experts.15.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.13.zaya_block.experts.local_experts.2.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.13.zaya_block.experts.local_experts.2.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.13.zaya_block.experts.local_experts.3.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.13.zaya_block.experts.local_experts.3.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.13.zaya_block.experts.local_experts.4.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.13.zaya_block.experts.local_experts.4.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.13.zaya_block.experts.local_experts.5.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.13.zaya_block.experts.local_experts.5.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.13.zaya_block.experts.local_experts.6.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.13.zaya_block.experts.local_experts.6.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.13.zaya_block.experts.local_experts.7.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.13.zaya_block.experts.local_experts.7.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.13.zaya_block.experts.local_experts.8.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.13.zaya_block.experts.local_experts.8.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.13.zaya_block.experts.local_experts.9.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.13.zaya_block.experts.local_experts.9.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.13.zaya_block.router.balancing_biases": "model-00001-of-00004.safetensors", + "model.layers.13.zaya_block.router.down_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.13.zaya_block.router.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.13.zaya_block.router.rmsnorm_eda.weight": "model-00001-of-00004.safetensors", + "model.layers.13.zaya_block.router.router_mlp.0.bias": "model-00001-of-00004.safetensors", + "model.layers.13.zaya_block.router.router_mlp.0.weight": "model-00001-of-00004.safetensors", + "model.layers.13.zaya_block.router.router_mlp.2.bias": "model-00001-of-00004.safetensors", + "model.layers.13.zaya_block.router.router_mlp.2.weight": "model-00001-of-00004.safetensors", + "model.layers.13.zaya_block.router.router_mlp.4.weight": "model-00001-of-00004.safetensors", + "model.layers.13.zaya_block.router.router_states_scale": "model-00001-of-00004.safetensors", + "model.layers.14.input_norm.weight": "model-00001-of-00004.safetensors", + "model.layers.14.res_scale.hidden_states_bias": "model-00001-of-00004.safetensors", + "model.layers.14.res_scale.hidden_states_scale": "model-00001-of-00004.safetensors", + "model.layers.14.res_scale.residual_bias": "model-00001-of-00004.safetensors", + "model.layers.14.res_scale.residual_scale": "model-00001-of-00004.safetensors", + "model.layers.14.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.14.self_attn.qkv.conv_qk.0.bias": "model-00001-of-00004.safetensors", + "model.layers.14.self_attn.qkv.conv_qk.0.weight": "model-00001-of-00004.safetensors", + "model.layers.14.self_attn.qkv.conv_qk.1.bias": "model-00001-of-00004.safetensors", + "model.layers.14.self_attn.qkv.conv_qk.1.weight": "model-00001-of-00004.safetensors", + "model.layers.14.self_attn.qkv.linear_k.weight": "model-00001-of-00004.safetensors", + "model.layers.14.self_attn.qkv.linear_q.weight": "model-00001-of-00004.safetensors", + "model.layers.14.self_attn.qkv.temp": "model-00001-of-00004.safetensors", + "model.layers.14.self_attn.qkv.val_proj1.weight": "model-00001-of-00004.safetensors", + "model.layers.14.self_attn.qkv.val_proj2.weight": "model-00001-of-00004.safetensors", + "model.layers.15.input_norm.weight": "model-00001-of-00004.safetensors", + "model.layers.15.res_scale.hidden_states_bias": "model-00001-of-00004.safetensors", + "model.layers.15.res_scale.hidden_states_scale": "model-00001-of-00004.safetensors", + "model.layers.15.res_scale.residual_bias": "model-00001-of-00004.safetensors", + "model.layers.15.res_scale.residual_scale": "model-00001-of-00004.safetensors", + "model.layers.15.zaya_block.experts.local_experts.0.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.15.zaya_block.experts.local_experts.0.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.15.zaya_block.experts.local_experts.1.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.15.zaya_block.experts.local_experts.1.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.15.zaya_block.experts.local_experts.10.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.15.zaya_block.experts.local_experts.10.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.15.zaya_block.experts.local_experts.11.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.15.zaya_block.experts.local_experts.11.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.15.zaya_block.experts.local_experts.12.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.15.zaya_block.experts.local_experts.12.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.15.zaya_block.experts.local_experts.13.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.15.zaya_block.experts.local_experts.13.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.15.zaya_block.experts.local_experts.14.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.15.zaya_block.experts.local_experts.14.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.15.zaya_block.experts.local_experts.15.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.15.zaya_block.experts.local_experts.15.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.15.zaya_block.experts.local_experts.2.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.15.zaya_block.experts.local_experts.2.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.15.zaya_block.experts.local_experts.3.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.15.zaya_block.experts.local_experts.3.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.15.zaya_block.experts.local_experts.4.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.15.zaya_block.experts.local_experts.4.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.15.zaya_block.experts.local_experts.5.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.15.zaya_block.experts.local_experts.5.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.15.zaya_block.experts.local_experts.6.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.15.zaya_block.experts.local_experts.6.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.15.zaya_block.experts.local_experts.7.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.15.zaya_block.experts.local_experts.7.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.15.zaya_block.experts.local_experts.8.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.15.zaya_block.experts.local_experts.8.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.15.zaya_block.experts.local_experts.9.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.15.zaya_block.experts.local_experts.9.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.15.zaya_block.router.balancing_biases": "model-00001-of-00004.safetensors", + "model.layers.15.zaya_block.router.down_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.15.zaya_block.router.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.15.zaya_block.router.rmsnorm_eda.weight": "model-00001-of-00004.safetensors", + "model.layers.15.zaya_block.router.router_mlp.0.bias": "model-00001-of-00004.safetensors", + "model.layers.15.zaya_block.router.router_mlp.0.weight": "model-00001-of-00004.safetensors", + "model.layers.15.zaya_block.router.router_mlp.2.bias": "model-00001-of-00004.safetensors", + "model.layers.15.zaya_block.router.router_mlp.2.weight": "model-00001-of-00004.safetensors", + "model.layers.15.zaya_block.router.router_mlp.4.weight": "model-00001-of-00004.safetensors", + "model.layers.15.zaya_block.router.router_states_scale": "model-00001-of-00004.safetensors", + "model.layers.16.input_norm.weight": "model-00001-of-00004.safetensors", + "model.layers.16.res_scale.hidden_states_bias": "model-00001-of-00004.safetensors", + "model.layers.16.res_scale.hidden_states_scale": "model-00001-of-00004.safetensors", + "model.layers.16.res_scale.residual_bias": "model-00001-of-00004.safetensors", + "model.layers.16.res_scale.residual_scale": "model-00001-of-00004.safetensors", + "model.layers.16.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.16.self_attn.qkv.conv_qk.0.bias": "model-00001-of-00004.safetensors", + "model.layers.16.self_attn.qkv.conv_qk.0.weight": "model-00001-of-00004.safetensors", + "model.layers.16.self_attn.qkv.conv_qk.1.bias": "model-00001-of-00004.safetensors", + "model.layers.16.self_attn.qkv.conv_qk.1.weight": "model-00001-of-00004.safetensors", + "model.layers.16.self_attn.qkv.linear_k.weight": "model-00001-of-00004.safetensors", + "model.layers.16.self_attn.qkv.linear_q.weight": "model-00001-of-00004.safetensors", + "model.layers.16.self_attn.qkv.temp": "model-00001-of-00004.safetensors", + "model.layers.16.self_attn.qkv.val_proj1.weight": "model-00001-of-00004.safetensors", + "model.layers.16.self_attn.qkv.val_proj2.weight": "model-00001-of-00004.safetensors", + "model.layers.17.input_norm.weight": "model-00001-of-00004.safetensors", + "model.layers.17.res_scale.hidden_states_bias": "model-00001-of-00004.safetensors", + "model.layers.17.res_scale.hidden_states_scale": "model-00001-of-00004.safetensors", + "model.layers.17.res_scale.residual_bias": "model-00001-of-00004.safetensors", + "model.layers.17.res_scale.residual_scale": "model-00001-of-00004.safetensors", + "model.layers.17.zaya_block.experts.local_experts.0.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.17.zaya_block.experts.local_experts.0.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.17.zaya_block.experts.local_experts.1.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.17.zaya_block.experts.local_experts.1.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.17.zaya_block.experts.local_experts.10.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.17.zaya_block.experts.local_experts.10.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.17.zaya_block.experts.local_experts.11.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.17.zaya_block.experts.local_experts.11.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.17.zaya_block.experts.local_experts.12.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.17.zaya_block.experts.local_experts.12.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.17.zaya_block.experts.local_experts.13.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.17.zaya_block.experts.local_experts.13.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.17.zaya_block.experts.local_experts.14.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.17.zaya_block.experts.local_experts.14.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.17.zaya_block.experts.local_experts.15.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.17.zaya_block.experts.local_experts.15.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.17.zaya_block.experts.local_experts.2.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.17.zaya_block.experts.local_experts.2.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.17.zaya_block.experts.local_experts.3.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.17.zaya_block.experts.local_experts.3.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.17.zaya_block.experts.local_experts.4.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.17.zaya_block.experts.local_experts.4.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.17.zaya_block.experts.local_experts.5.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.17.zaya_block.experts.local_experts.5.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.17.zaya_block.experts.local_experts.6.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.17.zaya_block.experts.local_experts.6.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.17.zaya_block.experts.local_experts.7.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.17.zaya_block.experts.local_experts.7.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.17.zaya_block.experts.local_experts.8.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.17.zaya_block.experts.local_experts.8.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.17.zaya_block.experts.local_experts.9.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.17.zaya_block.experts.local_experts.9.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.17.zaya_block.router.balancing_biases": "model-00001-of-00004.safetensors", + "model.layers.17.zaya_block.router.down_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.17.zaya_block.router.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.17.zaya_block.router.rmsnorm_eda.weight": "model-00001-of-00004.safetensors", + "model.layers.17.zaya_block.router.router_mlp.0.bias": "model-00001-of-00004.safetensors", + "model.layers.17.zaya_block.router.router_mlp.0.weight": "model-00001-of-00004.safetensors", + "model.layers.17.zaya_block.router.router_mlp.2.bias": "model-00001-of-00004.safetensors", + "model.layers.17.zaya_block.router.router_mlp.2.weight": "model-00001-of-00004.safetensors", + "model.layers.17.zaya_block.router.router_mlp.4.weight": "model-00001-of-00004.safetensors", + "model.layers.17.zaya_block.router.router_states_scale": "model-00001-of-00004.safetensors", + "model.layers.18.input_norm.weight": "model-00001-of-00004.safetensors", + "model.layers.18.res_scale.hidden_states_bias": "model-00001-of-00004.safetensors", + "model.layers.18.res_scale.hidden_states_scale": "model-00001-of-00004.safetensors", + "model.layers.18.res_scale.residual_bias": "model-00001-of-00004.safetensors", + "model.layers.18.res_scale.residual_scale": "model-00001-of-00004.safetensors", + "model.layers.18.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.18.self_attn.qkv.conv_qk.0.bias": "model-00001-of-00004.safetensors", + "model.layers.18.self_attn.qkv.conv_qk.0.weight": "model-00001-of-00004.safetensors", + "model.layers.18.self_attn.qkv.conv_qk.1.bias": "model-00001-of-00004.safetensors", + "model.layers.18.self_attn.qkv.conv_qk.1.weight": "model-00001-of-00004.safetensors", + "model.layers.18.self_attn.qkv.linear_k.weight": "model-00001-of-00004.safetensors", + "model.layers.18.self_attn.qkv.linear_q.weight": "model-00001-of-00004.safetensors", + "model.layers.18.self_attn.qkv.temp": "model-00001-of-00004.safetensors", + "model.layers.18.self_attn.qkv.val_proj1.weight": "model-00001-of-00004.safetensors", + "model.layers.18.self_attn.qkv.val_proj2.weight": "model-00001-of-00004.safetensors", + "model.layers.19.input_norm.weight": "model-00002-of-00004.safetensors", + "model.layers.19.res_scale.hidden_states_bias": "model-00002-of-00004.safetensors", + "model.layers.19.res_scale.hidden_states_scale": "model-00002-of-00004.safetensors", + "model.layers.19.res_scale.residual_bias": "model-00002-of-00004.safetensors", + "model.layers.19.res_scale.residual_scale": "model-00002-of-00004.safetensors", + "model.layers.19.zaya_block.experts.local_experts.0.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.19.zaya_block.experts.local_experts.0.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.19.zaya_block.experts.local_experts.1.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.19.zaya_block.experts.local_experts.1.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.19.zaya_block.experts.local_experts.10.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.19.zaya_block.experts.local_experts.10.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.19.zaya_block.experts.local_experts.11.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.19.zaya_block.experts.local_experts.11.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.19.zaya_block.experts.local_experts.12.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.19.zaya_block.experts.local_experts.12.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.19.zaya_block.experts.local_experts.13.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.19.zaya_block.experts.local_experts.13.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.19.zaya_block.experts.local_experts.14.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.19.zaya_block.experts.local_experts.14.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.19.zaya_block.experts.local_experts.15.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.19.zaya_block.experts.local_experts.15.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.19.zaya_block.experts.local_experts.2.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.19.zaya_block.experts.local_experts.2.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.19.zaya_block.experts.local_experts.3.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.19.zaya_block.experts.local_experts.3.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.19.zaya_block.experts.local_experts.4.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.19.zaya_block.experts.local_experts.4.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.19.zaya_block.experts.local_experts.5.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.19.zaya_block.experts.local_experts.5.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.19.zaya_block.experts.local_experts.6.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.19.zaya_block.experts.local_experts.6.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.19.zaya_block.experts.local_experts.7.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.19.zaya_block.experts.local_experts.7.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.19.zaya_block.experts.local_experts.8.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.19.zaya_block.experts.local_experts.8.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.19.zaya_block.experts.local_experts.9.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.19.zaya_block.experts.local_experts.9.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.19.zaya_block.router.balancing_biases": "model-00001-of-00004.safetensors", + "model.layers.19.zaya_block.router.down_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.19.zaya_block.router.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.19.zaya_block.router.rmsnorm_eda.weight": "model-00001-of-00004.safetensors", + "model.layers.19.zaya_block.router.router_mlp.0.bias": "model-00001-of-00004.safetensors", + "model.layers.19.zaya_block.router.router_mlp.0.weight": "model-00001-of-00004.safetensors", + "model.layers.19.zaya_block.router.router_mlp.2.bias": "model-00001-of-00004.safetensors", + "model.layers.19.zaya_block.router.router_mlp.2.weight": "model-00001-of-00004.safetensors", + "model.layers.19.zaya_block.router.router_mlp.4.weight": "model-00001-of-00004.safetensors", + "model.layers.19.zaya_block.router.router_states_scale": "model-00001-of-00004.safetensors", + "model.layers.2.input_norm.weight": "model-00001-of-00004.safetensors", + "model.layers.2.res_scale.hidden_states_bias": "model-00001-of-00004.safetensors", + "model.layers.2.res_scale.hidden_states_scale": "model-00001-of-00004.safetensors", + "model.layers.2.res_scale.residual_bias": "model-00001-of-00004.safetensors", + "model.layers.2.res_scale.residual_scale": "model-00001-of-00004.safetensors", + "model.layers.2.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.2.self_attn.qkv.conv_qk.0.bias": "model-00001-of-00004.safetensors", + "model.layers.2.self_attn.qkv.conv_qk.0.weight": "model-00001-of-00004.safetensors", + "model.layers.2.self_attn.qkv.conv_qk.1.bias": "model-00001-of-00004.safetensors", + "model.layers.2.self_attn.qkv.conv_qk.1.weight": "model-00001-of-00004.safetensors", + "model.layers.2.self_attn.qkv.linear_k.weight": "model-00001-of-00004.safetensors", + "model.layers.2.self_attn.qkv.linear_q.weight": "model-00001-of-00004.safetensors", + "model.layers.2.self_attn.qkv.temp": "model-00001-of-00004.safetensors", + "model.layers.2.self_attn.qkv.val_proj1.weight": "model-00001-of-00004.safetensors", + "model.layers.2.self_attn.qkv.val_proj2.weight": "model-00001-of-00004.safetensors", + "model.layers.20.input_norm.weight": "model-00002-of-00004.safetensors", + "model.layers.20.res_scale.hidden_states_bias": "model-00002-of-00004.safetensors", + "model.layers.20.res_scale.hidden_states_scale": "model-00002-of-00004.safetensors", + "model.layers.20.res_scale.residual_bias": "model-00002-of-00004.safetensors", + "model.layers.20.res_scale.residual_scale": "model-00002-of-00004.safetensors", + "model.layers.20.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.20.self_attn.qkv.conv_qk.0.bias": "model-00002-of-00004.safetensors", + "model.layers.20.self_attn.qkv.conv_qk.0.weight": "model-00002-of-00004.safetensors", + "model.layers.20.self_attn.qkv.conv_qk.1.bias": "model-00002-of-00004.safetensors", + "model.layers.20.self_attn.qkv.conv_qk.1.weight": "model-00002-of-00004.safetensors", + "model.layers.20.self_attn.qkv.linear_k.weight": "model-00002-of-00004.safetensors", + "model.layers.20.self_attn.qkv.linear_q.weight": "model-00002-of-00004.safetensors", + "model.layers.20.self_attn.qkv.temp": "model-00002-of-00004.safetensors", + "model.layers.20.self_attn.qkv.val_proj1.weight": "model-00002-of-00004.safetensors", + "model.layers.20.self_attn.qkv.val_proj2.weight": "model-00002-of-00004.safetensors", + "model.layers.21.input_norm.weight": "model-00002-of-00004.safetensors", + "model.layers.21.res_scale.hidden_states_bias": "model-00002-of-00004.safetensors", + "model.layers.21.res_scale.hidden_states_scale": "model-00002-of-00004.safetensors", + "model.layers.21.res_scale.residual_bias": "model-00002-of-00004.safetensors", + "model.layers.21.res_scale.residual_scale": "model-00002-of-00004.safetensors", + "model.layers.21.zaya_block.experts.local_experts.0.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.21.zaya_block.experts.local_experts.0.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.21.zaya_block.experts.local_experts.1.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.21.zaya_block.experts.local_experts.1.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.21.zaya_block.experts.local_experts.10.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.21.zaya_block.experts.local_experts.10.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.21.zaya_block.experts.local_experts.11.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.21.zaya_block.experts.local_experts.11.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.21.zaya_block.experts.local_experts.12.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.21.zaya_block.experts.local_experts.12.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.21.zaya_block.experts.local_experts.13.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.21.zaya_block.experts.local_experts.13.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.21.zaya_block.experts.local_experts.14.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.21.zaya_block.experts.local_experts.14.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.21.zaya_block.experts.local_experts.15.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.21.zaya_block.experts.local_experts.15.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.21.zaya_block.experts.local_experts.2.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.21.zaya_block.experts.local_experts.2.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.21.zaya_block.experts.local_experts.3.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.21.zaya_block.experts.local_experts.3.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.21.zaya_block.experts.local_experts.4.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.21.zaya_block.experts.local_experts.4.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.21.zaya_block.experts.local_experts.5.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.21.zaya_block.experts.local_experts.5.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.21.zaya_block.experts.local_experts.6.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.21.zaya_block.experts.local_experts.6.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.21.zaya_block.experts.local_experts.7.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.21.zaya_block.experts.local_experts.7.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.21.zaya_block.experts.local_experts.8.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.21.zaya_block.experts.local_experts.8.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.21.zaya_block.experts.local_experts.9.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.21.zaya_block.experts.local_experts.9.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.21.zaya_block.router.balancing_biases": "model-00002-of-00004.safetensors", + "model.layers.21.zaya_block.router.down_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.21.zaya_block.router.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.21.zaya_block.router.rmsnorm_eda.weight": "model-00002-of-00004.safetensors", + "model.layers.21.zaya_block.router.router_mlp.0.bias": "model-00002-of-00004.safetensors", + "model.layers.21.zaya_block.router.router_mlp.0.weight": "model-00002-of-00004.safetensors", + "model.layers.21.zaya_block.router.router_mlp.2.bias": "model-00002-of-00004.safetensors", + "model.layers.21.zaya_block.router.router_mlp.2.weight": "model-00002-of-00004.safetensors", + "model.layers.21.zaya_block.router.router_mlp.4.weight": "model-00002-of-00004.safetensors", + "model.layers.21.zaya_block.router.router_states_scale": "model-00002-of-00004.safetensors", + "model.layers.22.input_norm.weight": "model-00002-of-00004.safetensors", + "model.layers.22.res_scale.hidden_states_bias": "model-00002-of-00004.safetensors", + "model.layers.22.res_scale.hidden_states_scale": "model-00002-of-00004.safetensors", + "model.layers.22.res_scale.residual_bias": "model-00002-of-00004.safetensors", + "model.layers.22.res_scale.residual_scale": "model-00002-of-00004.safetensors", + "model.layers.22.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.22.self_attn.qkv.conv_qk.0.bias": "model-00002-of-00004.safetensors", + "model.layers.22.self_attn.qkv.conv_qk.0.weight": "model-00002-of-00004.safetensors", + "model.layers.22.self_attn.qkv.conv_qk.1.bias": "model-00002-of-00004.safetensors", + "model.layers.22.self_attn.qkv.conv_qk.1.weight": "model-00002-of-00004.safetensors", + "model.layers.22.self_attn.qkv.linear_k.weight": "model-00002-of-00004.safetensors", + "model.layers.22.self_attn.qkv.linear_q.weight": "model-00002-of-00004.safetensors", + "model.layers.22.self_attn.qkv.temp": "model-00002-of-00004.safetensors", + "model.layers.22.self_attn.qkv.val_proj1.weight": "model-00002-of-00004.safetensors", + "model.layers.22.self_attn.qkv.val_proj2.weight": "model-00002-of-00004.safetensors", + "model.layers.23.input_norm.weight": "model-00002-of-00004.safetensors", + "model.layers.23.res_scale.hidden_states_bias": "model-00002-of-00004.safetensors", + "model.layers.23.res_scale.hidden_states_scale": "model-00002-of-00004.safetensors", + "model.layers.23.res_scale.residual_bias": "model-00002-of-00004.safetensors", + "model.layers.23.res_scale.residual_scale": "model-00002-of-00004.safetensors", + "model.layers.23.zaya_block.experts.local_experts.0.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.23.zaya_block.experts.local_experts.0.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.23.zaya_block.experts.local_experts.1.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.23.zaya_block.experts.local_experts.1.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.23.zaya_block.experts.local_experts.10.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.23.zaya_block.experts.local_experts.10.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.23.zaya_block.experts.local_experts.11.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.23.zaya_block.experts.local_experts.11.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.23.zaya_block.experts.local_experts.12.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.23.zaya_block.experts.local_experts.12.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.23.zaya_block.experts.local_experts.13.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.23.zaya_block.experts.local_experts.13.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.23.zaya_block.experts.local_experts.14.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.23.zaya_block.experts.local_experts.14.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.23.zaya_block.experts.local_experts.15.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.23.zaya_block.experts.local_experts.15.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.23.zaya_block.experts.local_experts.2.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.23.zaya_block.experts.local_experts.2.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.23.zaya_block.experts.local_experts.3.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.23.zaya_block.experts.local_experts.3.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.23.zaya_block.experts.local_experts.4.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.23.zaya_block.experts.local_experts.4.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.23.zaya_block.experts.local_experts.5.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.23.zaya_block.experts.local_experts.5.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.23.zaya_block.experts.local_experts.6.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.23.zaya_block.experts.local_experts.6.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.23.zaya_block.experts.local_experts.7.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.23.zaya_block.experts.local_experts.7.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.23.zaya_block.experts.local_experts.8.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.23.zaya_block.experts.local_experts.8.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.23.zaya_block.experts.local_experts.9.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.23.zaya_block.experts.local_experts.9.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.23.zaya_block.router.balancing_biases": "model-00002-of-00004.safetensors", + "model.layers.23.zaya_block.router.down_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.23.zaya_block.router.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.23.zaya_block.router.rmsnorm_eda.weight": "model-00002-of-00004.safetensors", + "model.layers.23.zaya_block.router.router_mlp.0.bias": "model-00002-of-00004.safetensors", + "model.layers.23.zaya_block.router.router_mlp.0.weight": "model-00002-of-00004.safetensors", + "model.layers.23.zaya_block.router.router_mlp.2.bias": "model-00002-of-00004.safetensors", + "model.layers.23.zaya_block.router.router_mlp.2.weight": "model-00002-of-00004.safetensors", + "model.layers.23.zaya_block.router.router_mlp.4.weight": "model-00002-of-00004.safetensors", + "model.layers.23.zaya_block.router.router_states_scale": "model-00002-of-00004.safetensors", + "model.layers.24.input_norm.weight": "model-00002-of-00004.safetensors", + "model.layers.24.res_scale.hidden_states_bias": "model-00002-of-00004.safetensors", + "model.layers.24.res_scale.hidden_states_scale": "model-00002-of-00004.safetensors", + "model.layers.24.res_scale.residual_bias": "model-00002-of-00004.safetensors", + "model.layers.24.res_scale.residual_scale": "model-00002-of-00004.safetensors", + "model.layers.24.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.24.self_attn.qkv.conv_qk.0.bias": "model-00002-of-00004.safetensors", + "model.layers.24.self_attn.qkv.conv_qk.0.weight": "model-00002-of-00004.safetensors", + "model.layers.24.self_attn.qkv.conv_qk.1.bias": "model-00002-of-00004.safetensors", + "model.layers.24.self_attn.qkv.conv_qk.1.weight": "model-00002-of-00004.safetensors", + "model.layers.24.self_attn.qkv.linear_k.weight": "model-00002-of-00004.safetensors", + "model.layers.24.self_attn.qkv.linear_q.weight": "model-00002-of-00004.safetensors", + "model.layers.24.self_attn.qkv.temp": "model-00002-of-00004.safetensors", + "model.layers.24.self_attn.qkv.val_proj1.weight": "model-00002-of-00004.safetensors", + "model.layers.24.self_attn.qkv.val_proj2.weight": "model-00002-of-00004.safetensors", + "model.layers.25.input_norm.weight": "model-00002-of-00004.safetensors", + "model.layers.25.res_scale.hidden_states_bias": "model-00002-of-00004.safetensors", + "model.layers.25.res_scale.hidden_states_scale": "model-00002-of-00004.safetensors", + "model.layers.25.res_scale.residual_bias": "model-00002-of-00004.safetensors", + "model.layers.25.res_scale.residual_scale": "model-00002-of-00004.safetensors", + "model.layers.25.zaya_block.experts.local_experts.0.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.25.zaya_block.experts.local_experts.0.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.25.zaya_block.experts.local_experts.1.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.25.zaya_block.experts.local_experts.1.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.25.zaya_block.experts.local_experts.10.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.25.zaya_block.experts.local_experts.10.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.25.zaya_block.experts.local_experts.11.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.25.zaya_block.experts.local_experts.11.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.25.zaya_block.experts.local_experts.12.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.25.zaya_block.experts.local_experts.12.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.25.zaya_block.experts.local_experts.13.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.25.zaya_block.experts.local_experts.13.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.25.zaya_block.experts.local_experts.14.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.25.zaya_block.experts.local_experts.14.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.25.zaya_block.experts.local_experts.15.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.25.zaya_block.experts.local_experts.15.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.25.zaya_block.experts.local_experts.2.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.25.zaya_block.experts.local_experts.2.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.25.zaya_block.experts.local_experts.3.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.25.zaya_block.experts.local_experts.3.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.25.zaya_block.experts.local_experts.4.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.25.zaya_block.experts.local_experts.4.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.25.zaya_block.experts.local_experts.5.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.25.zaya_block.experts.local_experts.5.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.25.zaya_block.experts.local_experts.6.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.25.zaya_block.experts.local_experts.6.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.25.zaya_block.experts.local_experts.7.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.25.zaya_block.experts.local_experts.7.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.25.zaya_block.experts.local_experts.8.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.25.zaya_block.experts.local_experts.8.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.25.zaya_block.experts.local_experts.9.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.25.zaya_block.experts.local_experts.9.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.25.zaya_block.router.balancing_biases": "model-00002-of-00004.safetensors", + "model.layers.25.zaya_block.router.down_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.25.zaya_block.router.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.25.zaya_block.router.rmsnorm_eda.weight": "model-00002-of-00004.safetensors", + "model.layers.25.zaya_block.router.router_mlp.0.bias": "model-00002-of-00004.safetensors", + "model.layers.25.zaya_block.router.router_mlp.0.weight": "model-00002-of-00004.safetensors", + "model.layers.25.zaya_block.router.router_mlp.2.bias": "model-00002-of-00004.safetensors", + "model.layers.25.zaya_block.router.router_mlp.2.weight": "model-00002-of-00004.safetensors", + "model.layers.25.zaya_block.router.router_mlp.4.weight": "model-00002-of-00004.safetensors", + "model.layers.25.zaya_block.router.router_states_scale": "model-00002-of-00004.safetensors", + "model.layers.26.input_norm.weight": "model-00002-of-00004.safetensors", + "model.layers.26.res_scale.hidden_states_bias": "model-00002-of-00004.safetensors", + "model.layers.26.res_scale.hidden_states_scale": "model-00002-of-00004.safetensors", + "model.layers.26.res_scale.residual_bias": "model-00002-of-00004.safetensors", + "model.layers.26.res_scale.residual_scale": "model-00002-of-00004.safetensors", + "model.layers.26.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.26.self_attn.qkv.conv_qk.0.bias": "model-00002-of-00004.safetensors", + "model.layers.26.self_attn.qkv.conv_qk.0.weight": "model-00002-of-00004.safetensors", + "model.layers.26.self_attn.qkv.conv_qk.1.bias": "model-00002-of-00004.safetensors", + "model.layers.26.self_attn.qkv.conv_qk.1.weight": "model-00002-of-00004.safetensors", + "model.layers.26.self_attn.qkv.linear_k.weight": "model-00002-of-00004.safetensors", + "model.layers.26.self_attn.qkv.linear_q.weight": "model-00002-of-00004.safetensors", + "model.layers.26.self_attn.qkv.temp": "model-00002-of-00004.safetensors", + "model.layers.26.self_attn.qkv.val_proj1.weight": "model-00002-of-00004.safetensors", + "model.layers.26.self_attn.qkv.val_proj2.weight": "model-00002-of-00004.safetensors", + "model.layers.27.input_norm.weight": "model-00002-of-00004.safetensors", + "model.layers.27.res_scale.hidden_states_bias": "model-00002-of-00004.safetensors", + "model.layers.27.res_scale.hidden_states_scale": "model-00002-of-00004.safetensors", + "model.layers.27.res_scale.residual_bias": "model-00002-of-00004.safetensors", + "model.layers.27.res_scale.residual_scale": "model-00002-of-00004.safetensors", + "model.layers.27.zaya_block.experts.local_experts.0.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.27.zaya_block.experts.local_experts.0.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.27.zaya_block.experts.local_experts.1.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.27.zaya_block.experts.local_experts.1.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.27.zaya_block.experts.local_experts.10.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.27.zaya_block.experts.local_experts.10.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.27.zaya_block.experts.local_experts.11.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.27.zaya_block.experts.local_experts.11.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.27.zaya_block.experts.local_experts.12.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.27.zaya_block.experts.local_experts.12.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.27.zaya_block.experts.local_experts.13.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.27.zaya_block.experts.local_experts.13.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.27.zaya_block.experts.local_experts.14.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.27.zaya_block.experts.local_experts.14.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.27.zaya_block.experts.local_experts.15.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.27.zaya_block.experts.local_experts.15.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.27.zaya_block.experts.local_experts.2.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.27.zaya_block.experts.local_experts.2.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.27.zaya_block.experts.local_experts.3.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.27.zaya_block.experts.local_experts.3.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.27.zaya_block.experts.local_experts.4.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.27.zaya_block.experts.local_experts.4.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.27.zaya_block.experts.local_experts.5.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.27.zaya_block.experts.local_experts.5.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.27.zaya_block.experts.local_experts.6.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.27.zaya_block.experts.local_experts.6.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.27.zaya_block.experts.local_experts.7.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.27.zaya_block.experts.local_experts.7.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.27.zaya_block.experts.local_experts.8.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.27.zaya_block.experts.local_experts.8.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.27.zaya_block.experts.local_experts.9.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.27.zaya_block.experts.local_experts.9.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.27.zaya_block.router.balancing_biases": "model-00002-of-00004.safetensors", + "model.layers.27.zaya_block.router.down_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.27.zaya_block.router.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.27.zaya_block.router.rmsnorm_eda.weight": "model-00002-of-00004.safetensors", + "model.layers.27.zaya_block.router.router_mlp.0.bias": "model-00002-of-00004.safetensors", + "model.layers.27.zaya_block.router.router_mlp.0.weight": "model-00002-of-00004.safetensors", + "model.layers.27.zaya_block.router.router_mlp.2.bias": "model-00002-of-00004.safetensors", + "model.layers.27.zaya_block.router.router_mlp.2.weight": "model-00002-of-00004.safetensors", + "model.layers.27.zaya_block.router.router_mlp.4.weight": "model-00002-of-00004.safetensors", + "model.layers.27.zaya_block.router.router_states_scale": "model-00002-of-00004.safetensors", + "model.layers.28.input_norm.weight": "model-00002-of-00004.safetensors", + "model.layers.28.res_scale.hidden_states_bias": "model-00002-of-00004.safetensors", + "model.layers.28.res_scale.hidden_states_scale": "model-00002-of-00004.safetensors", + "model.layers.28.res_scale.residual_bias": "model-00002-of-00004.safetensors", + "model.layers.28.res_scale.residual_scale": "model-00002-of-00004.safetensors", + "model.layers.28.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.28.self_attn.qkv.conv_qk.0.bias": "model-00002-of-00004.safetensors", + "model.layers.28.self_attn.qkv.conv_qk.0.weight": "model-00002-of-00004.safetensors", + "model.layers.28.self_attn.qkv.conv_qk.1.bias": "model-00002-of-00004.safetensors", + "model.layers.28.self_attn.qkv.conv_qk.1.weight": "model-00002-of-00004.safetensors", + "model.layers.28.self_attn.qkv.linear_k.weight": "model-00002-of-00004.safetensors", + "model.layers.28.self_attn.qkv.linear_q.weight": "model-00002-of-00004.safetensors", + "model.layers.28.self_attn.qkv.temp": "model-00002-of-00004.safetensors", + "model.layers.28.self_attn.qkv.val_proj1.weight": "model-00002-of-00004.safetensors", + "model.layers.28.self_attn.qkv.val_proj2.weight": "model-00002-of-00004.safetensors", + "model.layers.29.input_norm.weight": "model-00002-of-00004.safetensors", + "model.layers.29.res_scale.hidden_states_bias": "model-00002-of-00004.safetensors", + "model.layers.29.res_scale.hidden_states_scale": "model-00002-of-00004.safetensors", + "model.layers.29.res_scale.residual_bias": "model-00002-of-00004.safetensors", + "model.layers.29.res_scale.residual_scale": "model-00002-of-00004.safetensors", + "model.layers.29.zaya_block.experts.local_experts.0.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.29.zaya_block.experts.local_experts.0.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.29.zaya_block.experts.local_experts.1.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.29.zaya_block.experts.local_experts.1.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.29.zaya_block.experts.local_experts.10.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.29.zaya_block.experts.local_experts.10.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.29.zaya_block.experts.local_experts.11.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.29.zaya_block.experts.local_experts.11.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.29.zaya_block.experts.local_experts.12.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.29.zaya_block.experts.local_experts.12.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.29.zaya_block.experts.local_experts.13.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.29.zaya_block.experts.local_experts.13.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.29.zaya_block.experts.local_experts.14.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.29.zaya_block.experts.local_experts.14.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.29.zaya_block.experts.local_experts.15.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.29.zaya_block.experts.local_experts.15.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.29.zaya_block.experts.local_experts.2.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.29.zaya_block.experts.local_experts.2.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.29.zaya_block.experts.local_experts.3.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.29.zaya_block.experts.local_experts.3.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.29.zaya_block.experts.local_experts.4.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.29.zaya_block.experts.local_experts.4.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.29.zaya_block.experts.local_experts.5.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.29.zaya_block.experts.local_experts.5.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.29.zaya_block.experts.local_experts.6.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.29.zaya_block.experts.local_experts.6.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.29.zaya_block.experts.local_experts.7.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.29.zaya_block.experts.local_experts.7.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.29.zaya_block.experts.local_experts.8.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.29.zaya_block.experts.local_experts.8.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.29.zaya_block.experts.local_experts.9.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.29.zaya_block.experts.local_experts.9.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.29.zaya_block.router.balancing_biases": "model-00002-of-00004.safetensors", + "model.layers.29.zaya_block.router.down_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.29.zaya_block.router.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.29.zaya_block.router.rmsnorm_eda.weight": "model-00002-of-00004.safetensors", + "model.layers.29.zaya_block.router.router_mlp.0.bias": "model-00002-of-00004.safetensors", + "model.layers.29.zaya_block.router.router_mlp.0.weight": "model-00002-of-00004.safetensors", + "model.layers.29.zaya_block.router.router_mlp.2.bias": "model-00002-of-00004.safetensors", + "model.layers.29.zaya_block.router.router_mlp.2.weight": "model-00002-of-00004.safetensors", + "model.layers.29.zaya_block.router.router_mlp.4.weight": "model-00002-of-00004.safetensors", + "model.layers.29.zaya_block.router.router_states_scale": "model-00002-of-00004.safetensors", + "model.layers.3.input_norm.weight": "model-00001-of-00004.safetensors", + "model.layers.3.res_scale.hidden_states_bias": "model-00001-of-00004.safetensors", + "model.layers.3.res_scale.hidden_states_scale": "model-00001-of-00004.safetensors", + "model.layers.3.res_scale.residual_bias": "model-00001-of-00004.safetensors", + "model.layers.3.res_scale.residual_scale": "model-00001-of-00004.safetensors", + "model.layers.3.zaya_block.experts.local_experts.0.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.3.zaya_block.experts.local_experts.0.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.3.zaya_block.experts.local_experts.1.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.3.zaya_block.experts.local_experts.1.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.3.zaya_block.experts.local_experts.10.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.3.zaya_block.experts.local_experts.10.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.3.zaya_block.experts.local_experts.11.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.3.zaya_block.experts.local_experts.11.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.3.zaya_block.experts.local_experts.12.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.3.zaya_block.experts.local_experts.12.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.3.zaya_block.experts.local_experts.13.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.3.zaya_block.experts.local_experts.13.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.3.zaya_block.experts.local_experts.14.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.3.zaya_block.experts.local_experts.14.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.3.zaya_block.experts.local_experts.15.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.3.zaya_block.experts.local_experts.15.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.3.zaya_block.experts.local_experts.2.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.3.zaya_block.experts.local_experts.2.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.3.zaya_block.experts.local_experts.3.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.3.zaya_block.experts.local_experts.3.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.3.zaya_block.experts.local_experts.4.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.3.zaya_block.experts.local_experts.4.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.3.zaya_block.experts.local_experts.5.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.3.zaya_block.experts.local_experts.5.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.3.zaya_block.experts.local_experts.6.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.3.zaya_block.experts.local_experts.6.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.3.zaya_block.experts.local_experts.7.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.3.zaya_block.experts.local_experts.7.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.3.zaya_block.experts.local_experts.8.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.3.zaya_block.experts.local_experts.8.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.3.zaya_block.experts.local_experts.9.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.3.zaya_block.experts.local_experts.9.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.3.zaya_block.router.balancing_biases": "model-00001-of-00004.safetensors", + "model.layers.3.zaya_block.router.down_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.3.zaya_block.router.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.3.zaya_block.router.rmsnorm_eda.weight": "model-00001-of-00004.safetensors", + "model.layers.3.zaya_block.router.router_mlp.0.bias": "model-00001-of-00004.safetensors", + "model.layers.3.zaya_block.router.router_mlp.0.weight": "model-00001-of-00004.safetensors", + "model.layers.3.zaya_block.router.router_mlp.2.bias": "model-00001-of-00004.safetensors", + "model.layers.3.zaya_block.router.router_mlp.2.weight": "model-00001-of-00004.safetensors", + "model.layers.3.zaya_block.router.router_mlp.4.weight": "model-00001-of-00004.safetensors", + "model.layers.3.zaya_block.router.router_states_scale": "model-00001-of-00004.safetensors", + "model.layers.30.input_norm.weight": "model-00002-of-00004.safetensors", + "model.layers.30.res_scale.hidden_states_bias": "model-00002-of-00004.safetensors", + "model.layers.30.res_scale.hidden_states_scale": "model-00002-of-00004.safetensors", + "model.layers.30.res_scale.residual_bias": "model-00002-of-00004.safetensors", + "model.layers.30.res_scale.residual_scale": "model-00002-of-00004.safetensors", + "model.layers.30.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.30.self_attn.qkv.conv_qk.0.bias": "model-00002-of-00004.safetensors", + "model.layers.30.self_attn.qkv.conv_qk.0.weight": "model-00002-of-00004.safetensors", + "model.layers.30.self_attn.qkv.conv_qk.1.bias": "model-00002-of-00004.safetensors", + "model.layers.30.self_attn.qkv.conv_qk.1.weight": "model-00002-of-00004.safetensors", + "model.layers.30.self_attn.qkv.linear_k.weight": "model-00002-of-00004.safetensors", + "model.layers.30.self_attn.qkv.linear_q.weight": "model-00002-of-00004.safetensors", + "model.layers.30.self_attn.qkv.temp": "model-00002-of-00004.safetensors", + "model.layers.30.self_attn.qkv.val_proj1.weight": "model-00002-of-00004.safetensors", + "model.layers.30.self_attn.qkv.val_proj2.weight": "model-00002-of-00004.safetensors", + "model.layers.31.input_norm.weight": "model-00002-of-00004.safetensors", + "model.layers.31.res_scale.hidden_states_bias": "model-00002-of-00004.safetensors", + "model.layers.31.res_scale.hidden_states_scale": "model-00002-of-00004.safetensors", + "model.layers.31.res_scale.residual_bias": "model-00002-of-00004.safetensors", + "model.layers.31.res_scale.residual_scale": "model-00002-of-00004.safetensors", + "model.layers.31.zaya_block.experts.local_experts.0.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.31.zaya_block.experts.local_experts.0.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.31.zaya_block.experts.local_experts.1.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.31.zaya_block.experts.local_experts.1.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.31.zaya_block.experts.local_experts.10.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.31.zaya_block.experts.local_experts.10.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.31.zaya_block.experts.local_experts.11.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.31.zaya_block.experts.local_experts.11.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.31.zaya_block.experts.local_experts.12.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.31.zaya_block.experts.local_experts.12.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.31.zaya_block.experts.local_experts.13.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.31.zaya_block.experts.local_experts.13.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.31.zaya_block.experts.local_experts.14.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.31.zaya_block.experts.local_experts.14.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.31.zaya_block.experts.local_experts.15.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.31.zaya_block.experts.local_experts.15.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.31.zaya_block.experts.local_experts.2.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.31.zaya_block.experts.local_experts.2.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.31.zaya_block.experts.local_experts.3.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.31.zaya_block.experts.local_experts.3.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.31.zaya_block.experts.local_experts.4.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.31.zaya_block.experts.local_experts.4.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.31.zaya_block.experts.local_experts.5.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.31.zaya_block.experts.local_experts.5.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.31.zaya_block.experts.local_experts.6.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.31.zaya_block.experts.local_experts.6.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.31.zaya_block.experts.local_experts.7.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.31.zaya_block.experts.local_experts.7.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.31.zaya_block.experts.local_experts.8.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.31.zaya_block.experts.local_experts.8.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.31.zaya_block.experts.local_experts.9.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.31.zaya_block.experts.local_experts.9.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.31.zaya_block.router.balancing_biases": "model-00002-of-00004.safetensors", + "model.layers.31.zaya_block.router.down_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.31.zaya_block.router.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.31.zaya_block.router.rmsnorm_eda.weight": "model-00002-of-00004.safetensors", + "model.layers.31.zaya_block.router.router_mlp.0.bias": "model-00002-of-00004.safetensors", + "model.layers.31.zaya_block.router.router_mlp.0.weight": "model-00002-of-00004.safetensors", + "model.layers.31.zaya_block.router.router_mlp.2.bias": "model-00002-of-00004.safetensors", + "model.layers.31.zaya_block.router.router_mlp.2.weight": "model-00002-of-00004.safetensors", + "model.layers.31.zaya_block.router.router_mlp.4.weight": "model-00002-of-00004.safetensors", + "model.layers.31.zaya_block.router.router_states_scale": "model-00002-of-00004.safetensors", + "model.layers.32.input_norm.weight": "model-00002-of-00004.safetensors", + "model.layers.32.res_scale.hidden_states_bias": "model-00002-of-00004.safetensors", + "model.layers.32.res_scale.hidden_states_scale": "model-00002-of-00004.safetensors", + "model.layers.32.res_scale.residual_bias": "model-00002-of-00004.safetensors", + "model.layers.32.res_scale.residual_scale": "model-00002-of-00004.safetensors", + "model.layers.32.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.32.self_attn.qkv.conv_qk.0.bias": "model-00002-of-00004.safetensors", + "model.layers.32.self_attn.qkv.conv_qk.0.weight": "model-00002-of-00004.safetensors", + "model.layers.32.self_attn.qkv.conv_qk.1.bias": "model-00002-of-00004.safetensors", + "model.layers.32.self_attn.qkv.conv_qk.1.weight": "model-00002-of-00004.safetensors", + "model.layers.32.self_attn.qkv.linear_k.weight": "model-00002-of-00004.safetensors", + "model.layers.32.self_attn.qkv.linear_q.weight": "model-00002-of-00004.safetensors", + "model.layers.32.self_attn.qkv.temp": "model-00002-of-00004.safetensors", + "model.layers.32.self_attn.qkv.val_proj1.weight": "model-00002-of-00004.safetensors", + "model.layers.32.self_attn.qkv.val_proj2.weight": "model-00002-of-00004.safetensors", + "model.layers.33.input_norm.weight": "model-00002-of-00004.safetensors", + "model.layers.33.res_scale.hidden_states_bias": "model-00002-of-00004.safetensors", + "model.layers.33.res_scale.hidden_states_scale": "model-00002-of-00004.safetensors", + "model.layers.33.res_scale.residual_bias": "model-00002-of-00004.safetensors", + "model.layers.33.res_scale.residual_scale": "model-00002-of-00004.safetensors", + "model.layers.33.zaya_block.experts.local_experts.0.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.33.zaya_block.experts.local_experts.0.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.33.zaya_block.experts.local_experts.1.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.33.zaya_block.experts.local_experts.1.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.33.zaya_block.experts.local_experts.10.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.33.zaya_block.experts.local_experts.10.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.33.zaya_block.experts.local_experts.11.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.33.zaya_block.experts.local_experts.11.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.33.zaya_block.experts.local_experts.12.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.33.zaya_block.experts.local_experts.12.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.33.zaya_block.experts.local_experts.13.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.33.zaya_block.experts.local_experts.13.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.33.zaya_block.experts.local_experts.14.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.33.zaya_block.experts.local_experts.14.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.33.zaya_block.experts.local_experts.15.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.33.zaya_block.experts.local_experts.15.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.33.zaya_block.experts.local_experts.2.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.33.zaya_block.experts.local_experts.2.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.33.zaya_block.experts.local_experts.3.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.33.zaya_block.experts.local_experts.3.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.33.zaya_block.experts.local_experts.4.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.33.zaya_block.experts.local_experts.4.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.33.zaya_block.experts.local_experts.5.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.33.zaya_block.experts.local_experts.5.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.33.zaya_block.experts.local_experts.6.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.33.zaya_block.experts.local_experts.6.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.33.zaya_block.experts.local_experts.7.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.33.zaya_block.experts.local_experts.7.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.33.zaya_block.experts.local_experts.8.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.33.zaya_block.experts.local_experts.8.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.33.zaya_block.experts.local_experts.9.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.33.zaya_block.experts.local_experts.9.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.33.zaya_block.router.balancing_biases": "model-00002-of-00004.safetensors", + "model.layers.33.zaya_block.router.down_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.33.zaya_block.router.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.33.zaya_block.router.rmsnorm_eda.weight": "model-00002-of-00004.safetensors", + "model.layers.33.zaya_block.router.router_mlp.0.bias": "model-00002-of-00004.safetensors", + "model.layers.33.zaya_block.router.router_mlp.0.weight": "model-00002-of-00004.safetensors", + "model.layers.33.zaya_block.router.router_mlp.2.bias": "model-00002-of-00004.safetensors", + "model.layers.33.zaya_block.router.router_mlp.2.weight": "model-00002-of-00004.safetensors", + "model.layers.33.zaya_block.router.router_mlp.4.weight": "model-00002-of-00004.safetensors", + "model.layers.33.zaya_block.router.router_states_scale": "model-00002-of-00004.safetensors", + "model.layers.34.input_norm.weight": "model-00002-of-00004.safetensors", + "model.layers.34.res_scale.hidden_states_bias": "model-00002-of-00004.safetensors", + "model.layers.34.res_scale.hidden_states_scale": "model-00002-of-00004.safetensors", + "model.layers.34.res_scale.residual_bias": "model-00002-of-00004.safetensors", + "model.layers.34.res_scale.residual_scale": "model-00002-of-00004.safetensors", + "model.layers.34.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.34.self_attn.qkv.conv_qk.0.bias": "model-00002-of-00004.safetensors", + "model.layers.34.self_attn.qkv.conv_qk.0.weight": "model-00002-of-00004.safetensors", + "model.layers.34.self_attn.qkv.conv_qk.1.bias": "model-00002-of-00004.safetensors", + "model.layers.34.self_attn.qkv.conv_qk.1.weight": "model-00002-of-00004.safetensors", + "model.layers.34.self_attn.qkv.linear_k.weight": "model-00002-of-00004.safetensors", + "model.layers.34.self_attn.qkv.linear_q.weight": "model-00002-of-00004.safetensors", + "model.layers.34.self_attn.qkv.temp": "model-00002-of-00004.safetensors", + "model.layers.34.self_attn.qkv.val_proj1.weight": "model-00002-of-00004.safetensors", + "model.layers.34.self_attn.qkv.val_proj2.weight": "model-00002-of-00004.safetensors", + "model.layers.35.input_norm.weight": "model-00002-of-00004.safetensors", + "model.layers.35.res_scale.hidden_states_bias": "model-00002-of-00004.safetensors", + "model.layers.35.res_scale.hidden_states_scale": "model-00002-of-00004.safetensors", + "model.layers.35.res_scale.residual_bias": "model-00002-of-00004.safetensors", + "model.layers.35.res_scale.residual_scale": "model-00002-of-00004.safetensors", + "model.layers.35.zaya_block.experts.local_experts.0.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.35.zaya_block.experts.local_experts.0.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.35.zaya_block.experts.local_experts.1.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.35.zaya_block.experts.local_experts.1.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.35.zaya_block.experts.local_experts.10.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.35.zaya_block.experts.local_experts.10.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.35.zaya_block.experts.local_experts.11.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.35.zaya_block.experts.local_experts.11.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.35.zaya_block.experts.local_experts.12.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.35.zaya_block.experts.local_experts.12.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.35.zaya_block.experts.local_experts.13.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.35.zaya_block.experts.local_experts.13.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.35.zaya_block.experts.local_experts.14.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.35.zaya_block.experts.local_experts.14.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.35.zaya_block.experts.local_experts.15.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.35.zaya_block.experts.local_experts.15.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.35.zaya_block.experts.local_experts.2.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.35.zaya_block.experts.local_experts.2.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.35.zaya_block.experts.local_experts.3.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.35.zaya_block.experts.local_experts.3.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.35.zaya_block.experts.local_experts.4.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.35.zaya_block.experts.local_experts.4.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.35.zaya_block.experts.local_experts.5.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.35.zaya_block.experts.local_experts.5.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.35.zaya_block.experts.local_experts.6.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.35.zaya_block.experts.local_experts.6.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.35.zaya_block.experts.local_experts.7.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.35.zaya_block.experts.local_experts.7.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.35.zaya_block.experts.local_experts.8.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.35.zaya_block.experts.local_experts.8.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.35.zaya_block.experts.local_experts.9.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.35.zaya_block.experts.local_experts.9.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.35.zaya_block.router.balancing_biases": "model-00002-of-00004.safetensors", + "model.layers.35.zaya_block.router.down_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.35.zaya_block.router.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.35.zaya_block.router.rmsnorm_eda.weight": "model-00002-of-00004.safetensors", + "model.layers.35.zaya_block.router.router_mlp.0.bias": "model-00002-of-00004.safetensors", + "model.layers.35.zaya_block.router.router_mlp.0.weight": "model-00002-of-00004.safetensors", + "model.layers.35.zaya_block.router.router_mlp.2.bias": "model-00002-of-00004.safetensors", + "model.layers.35.zaya_block.router.router_mlp.2.weight": "model-00002-of-00004.safetensors", + "model.layers.35.zaya_block.router.router_mlp.4.weight": "model-00002-of-00004.safetensors", + "model.layers.35.zaya_block.router.router_states_scale": "model-00002-of-00004.safetensors", + "model.layers.36.input_norm.weight": "model-00002-of-00004.safetensors", + "model.layers.36.res_scale.hidden_states_bias": "model-00002-of-00004.safetensors", + "model.layers.36.res_scale.hidden_states_scale": "model-00002-of-00004.safetensors", + "model.layers.36.res_scale.residual_bias": "model-00002-of-00004.safetensors", + "model.layers.36.res_scale.residual_scale": "model-00002-of-00004.safetensors", + "model.layers.36.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.36.self_attn.qkv.conv_qk.0.bias": "model-00002-of-00004.safetensors", + "model.layers.36.self_attn.qkv.conv_qk.0.weight": "model-00002-of-00004.safetensors", + "model.layers.36.self_attn.qkv.conv_qk.1.bias": "model-00002-of-00004.safetensors", + "model.layers.36.self_attn.qkv.conv_qk.1.weight": "model-00002-of-00004.safetensors", + "model.layers.36.self_attn.qkv.linear_k.weight": "model-00002-of-00004.safetensors", + "model.layers.36.self_attn.qkv.linear_q.weight": "model-00002-of-00004.safetensors", + "model.layers.36.self_attn.qkv.temp": "model-00002-of-00004.safetensors", + "model.layers.36.self_attn.qkv.val_proj1.weight": "model-00002-of-00004.safetensors", + "model.layers.36.self_attn.qkv.val_proj2.weight": "model-00002-of-00004.safetensors", + "model.layers.37.input_norm.weight": "model-00002-of-00004.safetensors", + "model.layers.37.res_scale.hidden_states_bias": "model-00002-of-00004.safetensors", + "model.layers.37.res_scale.hidden_states_scale": "model-00002-of-00004.safetensors", + "model.layers.37.res_scale.residual_bias": "model-00002-of-00004.safetensors", + "model.layers.37.res_scale.residual_scale": "model-00002-of-00004.safetensors", + "model.layers.37.zaya_block.experts.local_experts.0.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.37.zaya_block.experts.local_experts.0.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.37.zaya_block.experts.local_experts.1.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.37.zaya_block.experts.local_experts.1.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.37.zaya_block.experts.local_experts.10.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.37.zaya_block.experts.local_experts.10.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.37.zaya_block.experts.local_experts.11.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.37.zaya_block.experts.local_experts.11.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.37.zaya_block.experts.local_experts.12.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.37.zaya_block.experts.local_experts.12.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.37.zaya_block.experts.local_experts.13.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.37.zaya_block.experts.local_experts.13.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.37.zaya_block.experts.local_experts.14.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.37.zaya_block.experts.local_experts.14.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.37.zaya_block.experts.local_experts.15.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.37.zaya_block.experts.local_experts.15.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.37.zaya_block.experts.local_experts.2.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.37.zaya_block.experts.local_experts.2.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.37.zaya_block.experts.local_experts.3.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.37.zaya_block.experts.local_experts.3.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.37.zaya_block.experts.local_experts.4.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.37.zaya_block.experts.local_experts.4.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.37.zaya_block.experts.local_experts.5.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.37.zaya_block.experts.local_experts.5.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.37.zaya_block.experts.local_experts.6.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.37.zaya_block.experts.local_experts.6.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.37.zaya_block.experts.local_experts.7.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.37.zaya_block.experts.local_experts.7.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.37.zaya_block.experts.local_experts.8.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.37.zaya_block.experts.local_experts.8.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.37.zaya_block.experts.local_experts.9.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.37.zaya_block.experts.local_experts.9.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.37.zaya_block.router.balancing_biases": "model-00002-of-00004.safetensors", + "model.layers.37.zaya_block.router.down_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.37.zaya_block.router.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.37.zaya_block.router.rmsnorm_eda.weight": "model-00002-of-00004.safetensors", + "model.layers.37.zaya_block.router.router_mlp.0.bias": "model-00002-of-00004.safetensors", + "model.layers.37.zaya_block.router.router_mlp.0.weight": "model-00002-of-00004.safetensors", + "model.layers.37.zaya_block.router.router_mlp.2.bias": "model-00002-of-00004.safetensors", + "model.layers.37.zaya_block.router.router_mlp.2.weight": "model-00002-of-00004.safetensors", + "model.layers.37.zaya_block.router.router_mlp.4.weight": "model-00002-of-00004.safetensors", + "model.layers.37.zaya_block.router.router_states_scale": "model-00002-of-00004.safetensors", + "model.layers.38.input_norm.weight": "model-00002-of-00004.safetensors", + "model.layers.38.res_scale.hidden_states_bias": "model-00002-of-00004.safetensors", + "model.layers.38.res_scale.hidden_states_scale": "model-00002-of-00004.safetensors", + "model.layers.38.res_scale.residual_bias": "model-00002-of-00004.safetensors", + "model.layers.38.res_scale.residual_scale": "model-00002-of-00004.safetensors", + "model.layers.38.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.38.self_attn.qkv.conv_qk.0.bias": "model-00002-of-00004.safetensors", + "model.layers.38.self_attn.qkv.conv_qk.0.weight": "model-00002-of-00004.safetensors", + "model.layers.38.self_attn.qkv.conv_qk.1.bias": "model-00002-of-00004.safetensors", + "model.layers.38.self_attn.qkv.conv_qk.1.weight": "model-00002-of-00004.safetensors", + "model.layers.38.self_attn.qkv.linear_k.weight": "model-00002-of-00004.safetensors", + "model.layers.38.self_attn.qkv.linear_q.weight": "model-00002-of-00004.safetensors", + "model.layers.38.self_attn.qkv.temp": "model-00002-of-00004.safetensors", + "model.layers.38.self_attn.qkv.val_proj1.weight": "model-00002-of-00004.safetensors", + "model.layers.38.self_attn.qkv.val_proj2.weight": "model-00002-of-00004.safetensors", + "model.layers.39.input_norm.weight": "model-00002-of-00004.safetensors", + "model.layers.39.res_scale.hidden_states_bias": "model-00002-of-00004.safetensors", + "model.layers.39.res_scale.hidden_states_scale": "model-00002-of-00004.safetensors", + "model.layers.39.res_scale.residual_bias": "model-00002-of-00004.safetensors", + "model.layers.39.res_scale.residual_scale": "model-00002-of-00004.safetensors", + "model.layers.39.zaya_block.experts.local_experts.0.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.39.zaya_block.experts.local_experts.0.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.39.zaya_block.experts.local_experts.1.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.39.zaya_block.experts.local_experts.1.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.39.zaya_block.experts.local_experts.10.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.39.zaya_block.experts.local_experts.10.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.39.zaya_block.experts.local_experts.11.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.39.zaya_block.experts.local_experts.11.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.39.zaya_block.experts.local_experts.12.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.39.zaya_block.experts.local_experts.12.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.39.zaya_block.experts.local_experts.13.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.39.zaya_block.experts.local_experts.13.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.39.zaya_block.experts.local_experts.14.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.39.zaya_block.experts.local_experts.14.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.39.zaya_block.experts.local_experts.15.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.39.zaya_block.experts.local_experts.15.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.39.zaya_block.experts.local_experts.2.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.39.zaya_block.experts.local_experts.2.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.39.zaya_block.experts.local_experts.3.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.39.zaya_block.experts.local_experts.3.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.39.zaya_block.experts.local_experts.4.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.39.zaya_block.experts.local_experts.4.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.39.zaya_block.experts.local_experts.5.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.39.zaya_block.experts.local_experts.5.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.39.zaya_block.experts.local_experts.6.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.39.zaya_block.experts.local_experts.6.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.39.zaya_block.experts.local_experts.7.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.39.zaya_block.experts.local_experts.7.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.39.zaya_block.experts.local_experts.8.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.39.zaya_block.experts.local_experts.8.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.39.zaya_block.experts.local_experts.9.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.39.zaya_block.experts.local_experts.9.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.39.zaya_block.router.balancing_biases": "model-00002-of-00004.safetensors", + "model.layers.39.zaya_block.router.down_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.39.zaya_block.router.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.39.zaya_block.router.rmsnorm_eda.weight": "model-00002-of-00004.safetensors", + "model.layers.39.zaya_block.router.router_mlp.0.bias": "model-00002-of-00004.safetensors", + "model.layers.39.zaya_block.router.router_mlp.0.weight": "model-00002-of-00004.safetensors", + "model.layers.39.zaya_block.router.router_mlp.2.bias": "model-00002-of-00004.safetensors", + "model.layers.39.zaya_block.router.router_mlp.2.weight": "model-00002-of-00004.safetensors", + "model.layers.39.zaya_block.router.router_mlp.4.weight": "model-00002-of-00004.safetensors", + "model.layers.39.zaya_block.router.router_states_scale": "model-00002-of-00004.safetensors", + "model.layers.4.input_norm.weight": "model-00001-of-00004.safetensors", + "model.layers.4.res_scale.hidden_states_bias": "model-00001-of-00004.safetensors", + "model.layers.4.res_scale.hidden_states_scale": "model-00001-of-00004.safetensors", + "model.layers.4.res_scale.residual_bias": "model-00001-of-00004.safetensors", + "model.layers.4.res_scale.residual_scale": "model-00001-of-00004.safetensors", + "model.layers.4.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.4.self_attn.qkv.conv_qk.0.bias": "model-00001-of-00004.safetensors", + "model.layers.4.self_attn.qkv.conv_qk.0.weight": "model-00001-of-00004.safetensors", + "model.layers.4.self_attn.qkv.conv_qk.1.bias": "model-00001-of-00004.safetensors", + "model.layers.4.self_attn.qkv.conv_qk.1.weight": "model-00001-of-00004.safetensors", + "model.layers.4.self_attn.qkv.linear_k.weight": "model-00001-of-00004.safetensors", + "model.layers.4.self_attn.qkv.linear_q.weight": "model-00001-of-00004.safetensors", + "model.layers.4.self_attn.qkv.temp": "model-00001-of-00004.safetensors", + "model.layers.4.self_attn.qkv.val_proj1.weight": "model-00001-of-00004.safetensors", + "model.layers.4.self_attn.qkv.val_proj2.weight": "model-00001-of-00004.safetensors", + "model.layers.40.input_norm.weight": "model-00002-of-00004.safetensors", + "model.layers.40.res_scale.hidden_states_bias": "model-00002-of-00004.safetensors", + "model.layers.40.res_scale.hidden_states_scale": "model-00002-of-00004.safetensors", + "model.layers.40.res_scale.residual_bias": "model-00002-of-00004.safetensors", + "model.layers.40.res_scale.residual_scale": "model-00002-of-00004.safetensors", + "model.layers.40.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.40.self_attn.qkv.conv_qk.0.bias": "model-00002-of-00004.safetensors", + "model.layers.40.self_attn.qkv.conv_qk.0.weight": "model-00002-of-00004.safetensors", + "model.layers.40.self_attn.qkv.conv_qk.1.bias": "model-00002-of-00004.safetensors", + "model.layers.40.self_attn.qkv.conv_qk.1.weight": "model-00002-of-00004.safetensors", + "model.layers.40.self_attn.qkv.linear_k.weight": "model-00002-of-00004.safetensors", + "model.layers.40.self_attn.qkv.linear_q.weight": "model-00002-of-00004.safetensors", + "model.layers.40.self_attn.qkv.temp": "model-00002-of-00004.safetensors", + "model.layers.40.self_attn.qkv.val_proj1.weight": "model-00002-of-00004.safetensors", + "model.layers.40.self_attn.qkv.val_proj2.weight": "model-00002-of-00004.safetensors", + "model.layers.41.input_norm.weight": "model-00002-of-00004.safetensors", + "model.layers.41.res_scale.hidden_states_bias": "model-00002-of-00004.safetensors", + "model.layers.41.res_scale.hidden_states_scale": "model-00002-of-00004.safetensors", + "model.layers.41.res_scale.residual_bias": "model-00002-of-00004.safetensors", + "model.layers.41.res_scale.residual_scale": "model-00002-of-00004.safetensors", + "model.layers.41.zaya_block.experts.local_experts.0.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.41.zaya_block.experts.local_experts.0.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.41.zaya_block.experts.local_experts.1.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.41.zaya_block.experts.local_experts.1.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.41.zaya_block.experts.local_experts.10.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.41.zaya_block.experts.local_experts.10.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.41.zaya_block.experts.local_experts.11.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.41.zaya_block.experts.local_experts.11.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.41.zaya_block.experts.local_experts.12.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.41.zaya_block.experts.local_experts.12.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.41.zaya_block.experts.local_experts.13.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.41.zaya_block.experts.local_experts.13.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.41.zaya_block.experts.local_experts.14.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.41.zaya_block.experts.local_experts.14.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.41.zaya_block.experts.local_experts.15.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.41.zaya_block.experts.local_experts.15.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.41.zaya_block.experts.local_experts.2.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.41.zaya_block.experts.local_experts.2.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.41.zaya_block.experts.local_experts.3.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.41.zaya_block.experts.local_experts.3.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.41.zaya_block.experts.local_experts.4.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.41.zaya_block.experts.local_experts.4.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.41.zaya_block.experts.local_experts.5.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.41.zaya_block.experts.local_experts.5.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.41.zaya_block.experts.local_experts.6.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.41.zaya_block.experts.local_experts.6.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.41.zaya_block.experts.local_experts.7.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.41.zaya_block.experts.local_experts.7.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.41.zaya_block.experts.local_experts.8.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.41.zaya_block.experts.local_experts.8.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.41.zaya_block.experts.local_experts.9.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.41.zaya_block.experts.local_experts.9.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.41.zaya_block.router.balancing_biases": "model-00002-of-00004.safetensors", + "model.layers.41.zaya_block.router.down_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.41.zaya_block.router.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.41.zaya_block.router.rmsnorm_eda.weight": "model-00002-of-00004.safetensors", + "model.layers.41.zaya_block.router.router_mlp.0.bias": "model-00002-of-00004.safetensors", + "model.layers.41.zaya_block.router.router_mlp.0.weight": "model-00002-of-00004.safetensors", + "model.layers.41.zaya_block.router.router_mlp.2.bias": "model-00002-of-00004.safetensors", + "model.layers.41.zaya_block.router.router_mlp.2.weight": "model-00002-of-00004.safetensors", + "model.layers.41.zaya_block.router.router_mlp.4.weight": "model-00002-of-00004.safetensors", + "model.layers.41.zaya_block.router.router_states_scale": "model-00002-of-00004.safetensors", + "model.layers.42.input_norm.weight": "model-00002-of-00004.safetensors", + "model.layers.42.res_scale.hidden_states_bias": "model-00002-of-00004.safetensors", + "model.layers.42.res_scale.hidden_states_scale": "model-00002-of-00004.safetensors", + "model.layers.42.res_scale.residual_bias": "model-00002-of-00004.safetensors", + "model.layers.42.res_scale.residual_scale": "model-00002-of-00004.safetensors", + "model.layers.42.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.42.self_attn.qkv.conv_qk.0.bias": "model-00002-of-00004.safetensors", + "model.layers.42.self_attn.qkv.conv_qk.0.weight": "model-00002-of-00004.safetensors", + "model.layers.42.self_attn.qkv.conv_qk.1.bias": "model-00002-of-00004.safetensors", + "model.layers.42.self_attn.qkv.conv_qk.1.weight": "model-00002-of-00004.safetensors", + "model.layers.42.self_attn.qkv.linear_k.weight": "model-00002-of-00004.safetensors", + "model.layers.42.self_attn.qkv.linear_q.weight": "model-00002-of-00004.safetensors", + "model.layers.42.self_attn.qkv.temp": "model-00002-of-00004.safetensors", + "model.layers.42.self_attn.qkv.val_proj1.weight": "model-00002-of-00004.safetensors", + "model.layers.42.self_attn.qkv.val_proj2.weight": "model-00002-of-00004.safetensors", + "model.layers.43.input_norm.weight": "model-00003-of-00004.safetensors", + "model.layers.43.res_scale.hidden_states_bias": "model-00003-of-00004.safetensors", + "model.layers.43.res_scale.hidden_states_scale": "model-00003-of-00004.safetensors", + "model.layers.43.res_scale.residual_bias": "model-00003-of-00004.safetensors", + "model.layers.43.res_scale.residual_scale": "model-00003-of-00004.safetensors", + "model.layers.43.zaya_block.experts.local_experts.0.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.43.zaya_block.experts.local_experts.0.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.43.zaya_block.experts.local_experts.1.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.43.zaya_block.experts.local_experts.1.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.43.zaya_block.experts.local_experts.10.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.43.zaya_block.experts.local_experts.10.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.43.zaya_block.experts.local_experts.11.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.43.zaya_block.experts.local_experts.11.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.43.zaya_block.experts.local_experts.12.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.43.zaya_block.experts.local_experts.12.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.43.zaya_block.experts.local_experts.13.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.43.zaya_block.experts.local_experts.13.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.43.zaya_block.experts.local_experts.14.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.43.zaya_block.experts.local_experts.14.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.43.zaya_block.experts.local_experts.15.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.43.zaya_block.experts.local_experts.15.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.43.zaya_block.experts.local_experts.2.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.43.zaya_block.experts.local_experts.2.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.43.zaya_block.experts.local_experts.3.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.43.zaya_block.experts.local_experts.3.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.43.zaya_block.experts.local_experts.4.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.43.zaya_block.experts.local_experts.4.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.43.zaya_block.experts.local_experts.5.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.43.zaya_block.experts.local_experts.5.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.43.zaya_block.experts.local_experts.6.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.43.zaya_block.experts.local_experts.6.linear_fc2.weight": "model-00002-of-00004.safetensors", + "model.layers.43.zaya_block.experts.local_experts.7.linear_fc1.weight": "model-00002-of-00004.safetensors", + "model.layers.43.zaya_block.experts.local_experts.7.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.43.zaya_block.experts.local_experts.8.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.43.zaya_block.experts.local_experts.8.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.43.zaya_block.experts.local_experts.9.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.43.zaya_block.experts.local_experts.9.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.43.zaya_block.router.balancing_biases": "model-00002-of-00004.safetensors", + "model.layers.43.zaya_block.router.down_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.43.zaya_block.router.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.43.zaya_block.router.rmsnorm_eda.weight": "model-00002-of-00004.safetensors", + "model.layers.43.zaya_block.router.router_mlp.0.bias": "model-00002-of-00004.safetensors", + "model.layers.43.zaya_block.router.router_mlp.0.weight": "model-00002-of-00004.safetensors", + "model.layers.43.zaya_block.router.router_mlp.2.bias": "model-00002-of-00004.safetensors", + "model.layers.43.zaya_block.router.router_mlp.2.weight": "model-00002-of-00004.safetensors", + "model.layers.43.zaya_block.router.router_mlp.4.weight": "model-00002-of-00004.safetensors", + "model.layers.43.zaya_block.router.router_states_scale": "model-00002-of-00004.safetensors", + "model.layers.44.input_norm.weight": "model-00003-of-00004.safetensors", + "model.layers.44.res_scale.hidden_states_bias": "model-00003-of-00004.safetensors", + "model.layers.44.res_scale.hidden_states_scale": "model-00003-of-00004.safetensors", + "model.layers.44.res_scale.residual_bias": "model-00003-of-00004.safetensors", + "model.layers.44.res_scale.residual_scale": "model-00003-of-00004.safetensors", + "model.layers.44.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.44.self_attn.qkv.conv_qk.0.bias": "model-00003-of-00004.safetensors", + "model.layers.44.self_attn.qkv.conv_qk.0.weight": "model-00003-of-00004.safetensors", + "model.layers.44.self_attn.qkv.conv_qk.1.bias": "model-00003-of-00004.safetensors", + "model.layers.44.self_attn.qkv.conv_qk.1.weight": "model-00003-of-00004.safetensors", + "model.layers.44.self_attn.qkv.linear_k.weight": "model-00003-of-00004.safetensors", + "model.layers.44.self_attn.qkv.linear_q.weight": "model-00003-of-00004.safetensors", + "model.layers.44.self_attn.qkv.temp": "model-00003-of-00004.safetensors", + "model.layers.44.self_attn.qkv.val_proj1.weight": "model-00003-of-00004.safetensors", + "model.layers.44.self_attn.qkv.val_proj2.weight": "model-00003-of-00004.safetensors", + "model.layers.45.input_norm.weight": "model-00003-of-00004.safetensors", + "model.layers.45.res_scale.hidden_states_bias": "model-00003-of-00004.safetensors", + "model.layers.45.res_scale.hidden_states_scale": "model-00003-of-00004.safetensors", + "model.layers.45.res_scale.residual_bias": "model-00003-of-00004.safetensors", + "model.layers.45.res_scale.residual_scale": "model-00003-of-00004.safetensors", + "model.layers.45.zaya_block.experts.local_experts.0.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.45.zaya_block.experts.local_experts.0.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.45.zaya_block.experts.local_experts.1.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.45.zaya_block.experts.local_experts.1.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.45.zaya_block.experts.local_experts.10.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.45.zaya_block.experts.local_experts.10.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.45.zaya_block.experts.local_experts.11.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.45.zaya_block.experts.local_experts.11.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.45.zaya_block.experts.local_experts.12.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.45.zaya_block.experts.local_experts.12.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.45.zaya_block.experts.local_experts.13.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.45.zaya_block.experts.local_experts.13.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.45.zaya_block.experts.local_experts.14.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.45.zaya_block.experts.local_experts.14.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.45.zaya_block.experts.local_experts.15.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.45.zaya_block.experts.local_experts.15.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.45.zaya_block.experts.local_experts.2.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.45.zaya_block.experts.local_experts.2.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.45.zaya_block.experts.local_experts.3.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.45.zaya_block.experts.local_experts.3.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.45.zaya_block.experts.local_experts.4.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.45.zaya_block.experts.local_experts.4.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.45.zaya_block.experts.local_experts.5.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.45.zaya_block.experts.local_experts.5.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.45.zaya_block.experts.local_experts.6.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.45.zaya_block.experts.local_experts.6.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.45.zaya_block.experts.local_experts.7.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.45.zaya_block.experts.local_experts.7.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.45.zaya_block.experts.local_experts.8.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.45.zaya_block.experts.local_experts.8.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.45.zaya_block.experts.local_experts.9.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.45.zaya_block.experts.local_experts.9.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.45.zaya_block.router.balancing_biases": "model-00003-of-00004.safetensors", + "model.layers.45.zaya_block.router.down_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.45.zaya_block.router.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.45.zaya_block.router.rmsnorm_eda.weight": "model-00003-of-00004.safetensors", + "model.layers.45.zaya_block.router.router_mlp.0.bias": "model-00003-of-00004.safetensors", + "model.layers.45.zaya_block.router.router_mlp.0.weight": "model-00003-of-00004.safetensors", + "model.layers.45.zaya_block.router.router_mlp.2.bias": "model-00003-of-00004.safetensors", + "model.layers.45.zaya_block.router.router_mlp.2.weight": "model-00003-of-00004.safetensors", + "model.layers.45.zaya_block.router.router_mlp.4.weight": "model-00003-of-00004.safetensors", + "model.layers.45.zaya_block.router.router_states_scale": "model-00003-of-00004.safetensors", + "model.layers.46.input_norm.weight": "model-00003-of-00004.safetensors", + "model.layers.46.res_scale.hidden_states_bias": "model-00003-of-00004.safetensors", + "model.layers.46.res_scale.hidden_states_scale": "model-00003-of-00004.safetensors", + "model.layers.46.res_scale.residual_bias": "model-00003-of-00004.safetensors", + "model.layers.46.res_scale.residual_scale": "model-00003-of-00004.safetensors", + "model.layers.46.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.46.self_attn.qkv.conv_qk.0.bias": "model-00003-of-00004.safetensors", + "model.layers.46.self_attn.qkv.conv_qk.0.weight": "model-00003-of-00004.safetensors", + "model.layers.46.self_attn.qkv.conv_qk.1.bias": "model-00003-of-00004.safetensors", + "model.layers.46.self_attn.qkv.conv_qk.1.weight": "model-00003-of-00004.safetensors", + "model.layers.46.self_attn.qkv.linear_k.weight": "model-00003-of-00004.safetensors", + "model.layers.46.self_attn.qkv.linear_q.weight": "model-00003-of-00004.safetensors", + "model.layers.46.self_attn.qkv.temp": "model-00003-of-00004.safetensors", + "model.layers.46.self_attn.qkv.val_proj1.weight": "model-00003-of-00004.safetensors", + "model.layers.46.self_attn.qkv.val_proj2.weight": "model-00003-of-00004.safetensors", + "model.layers.47.input_norm.weight": "model-00003-of-00004.safetensors", + "model.layers.47.res_scale.hidden_states_bias": "model-00003-of-00004.safetensors", + "model.layers.47.res_scale.hidden_states_scale": "model-00003-of-00004.safetensors", + "model.layers.47.res_scale.residual_bias": "model-00003-of-00004.safetensors", + "model.layers.47.res_scale.residual_scale": "model-00003-of-00004.safetensors", + "model.layers.47.zaya_block.experts.local_experts.0.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.47.zaya_block.experts.local_experts.0.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.47.zaya_block.experts.local_experts.1.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.47.zaya_block.experts.local_experts.1.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.47.zaya_block.experts.local_experts.10.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.47.zaya_block.experts.local_experts.10.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.47.zaya_block.experts.local_experts.11.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.47.zaya_block.experts.local_experts.11.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.47.zaya_block.experts.local_experts.12.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.47.zaya_block.experts.local_experts.12.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.47.zaya_block.experts.local_experts.13.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.47.zaya_block.experts.local_experts.13.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.47.zaya_block.experts.local_experts.14.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.47.zaya_block.experts.local_experts.14.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.47.zaya_block.experts.local_experts.15.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.47.zaya_block.experts.local_experts.15.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.47.zaya_block.experts.local_experts.2.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.47.zaya_block.experts.local_experts.2.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.47.zaya_block.experts.local_experts.3.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.47.zaya_block.experts.local_experts.3.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.47.zaya_block.experts.local_experts.4.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.47.zaya_block.experts.local_experts.4.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.47.zaya_block.experts.local_experts.5.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.47.zaya_block.experts.local_experts.5.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.47.zaya_block.experts.local_experts.6.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.47.zaya_block.experts.local_experts.6.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.47.zaya_block.experts.local_experts.7.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.47.zaya_block.experts.local_experts.7.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.47.zaya_block.experts.local_experts.8.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.47.zaya_block.experts.local_experts.8.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.47.zaya_block.experts.local_experts.9.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.47.zaya_block.experts.local_experts.9.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.47.zaya_block.router.balancing_biases": "model-00003-of-00004.safetensors", + "model.layers.47.zaya_block.router.down_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.47.zaya_block.router.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.47.zaya_block.router.rmsnorm_eda.weight": "model-00003-of-00004.safetensors", + "model.layers.47.zaya_block.router.router_mlp.0.bias": "model-00003-of-00004.safetensors", + "model.layers.47.zaya_block.router.router_mlp.0.weight": "model-00003-of-00004.safetensors", + "model.layers.47.zaya_block.router.router_mlp.2.bias": "model-00003-of-00004.safetensors", + "model.layers.47.zaya_block.router.router_mlp.2.weight": "model-00003-of-00004.safetensors", + "model.layers.47.zaya_block.router.router_mlp.4.weight": "model-00003-of-00004.safetensors", + "model.layers.47.zaya_block.router.router_states_scale": "model-00003-of-00004.safetensors", + "model.layers.48.input_norm.weight": "model-00003-of-00004.safetensors", + "model.layers.48.res_scale.hidden_states_bias": "model-00003-of-00004.safetensors", + "model.layers.48.res_scale.hidden_states_scale": "model-00003-of-00004.safetensors", + "model.layers.48.res_scale.residual_bias": "model-00003-of-00004.safetensors", + "model.layers.48.res_scale.residual_scale": "model-00003-of-00004.safetensors", + "model.layers.48.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.48.self_attn.qkv.conv_qk.0.bias": "model-00003-of-00004.safetensors", + "model.layers.48.self_attn.qkv.conv_qk.0.weight": "model-00003-of-00004.safetensors", + "model.layers.48.self_attn.qkv.conv_qk.1.bias": "model-00003-of-00004.safetensors", + "model.layers.48.self_attn.qkv.conv_qk.1.weight": "model-00003-of-00004.safetensors", + "model.layers.48.self_attn.qkv.linear_k.weight": "model-00003-of-00004.safetensors", + "model.layers.48.self_attn.qkv.linear_q.weight": "model-00003-of-00004.safetensors", + "model.layers.48.self_attn.qkv.temp": "model-00003-of-00004.safetensors", + "model.layers.48.self_attn.qkv.val_proj1.weight": "model-00003-of-00004.safetensors", + "model.layers.48.self_attn.qkv.val_proj2.weight": "model-00003-of-00004.safetensors", + "model.layers.49.input_norm.weight": "model-00003-of-00004.safetensors", + "model.layers.49.res_scale.hidden_states_bias": "model-00003-of-00004.safetensors", + "model.layers.49.res_scale.hidden_states_scale": "model-00003-of-00004.safetensors", + "model.layers.49.res_scale.residual_bias": "model-00003-of-00004.safetensors", + "model.layers.49.res_scale.residual_scale": "model-00003-of-00004.safetensors", + "model.layers.49.zaya_block.experts.local_experts.0.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.49.zaya_block.experts.local_experts.0.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.49.zaya_block.experts.local_experts.1.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.49.zaya_block.experts.local_experts.1.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.49.zaya_block.experts.local_experts.10.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.49.zaya_block.experts.local_experts.10.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.49.zaya_block.experts.local_experts.11.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.49.zaya_block.experts.local_experts.11.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.49.zaya_block.experts.local_experts.12.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.49.zaya_block.experts.local_experts.12.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.49.zaya_block.experts.local_experts.13.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.49.zaya_block.experts.local_experts.13.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.49.zaya_block.experts.local_experts.14.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.49.zaya_block.experts.local_experts.14.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.49.zaya_block.experts.local_experts.15.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.49.zaya_block.experts.local_experts.15.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.49.zaya_block.experts.local_experts.2.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.49.zaya_block.experts.local_experts.2.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.49.zaya_block.experts.local_experts.3.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.49.zaya_block.experts.local_experts.3.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.49.zaya_block.experts.local_experts.4.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.49.zaya_block.experts.local_experts.4.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.49.zaya_block.experts.local_experts.5.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.49.zaya_block.experts.local_experts.5.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.49.zaya_block.experts.local_experts.6.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.49.zaya_block.experts.local_experts.6.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.49.zaya_block.experts.local_experts.7.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.49.zaya_block.experts.local_experts.7.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.49.zaya_block.experts.local_experts.8.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.49.zaya_block.experts.local_experts.8.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.49.zaya_block.experts.local_experts.9.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.49.zaya_block.experts.local_experts.9.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.49.zaya_block.router.balancing_biases": "model-00003-of-00004.safetensors", + "model.layers.49.zaya_block.router.down_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.49.zaya_block.router.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.49.zaya_block.router.rmsnorm_eda.weight": "model-00003-of-00004.safetensors", + "model.layers.49.zaya_block.router.router_mlp.0.bias": "model-00003-of-00004.safetensors", + "model.layers.49.zaya_block.router.router_mlp.0.weight": "model-00003-of-00004.safetensors", + "model.layers.49.zaya_block.router.router_mlp.2.bias": "model-00003-of-00004.safetensors", + "model.layers.49.zaya_block.router.router_mlp.2.weight": "model-00003-of-00004.safetensors", + "model.layers.49.zaya_block.router.router_mlp.4.weight": "model-00003-of-00004.safetensors", + "model.layers.49.zaya_block.router.router_states_scale": "model-00003-of-00004.safetensors", + "model.layers.5.input_norm.weight": "model-00001-of-00004.safetensors", + "model.layers.5.res_scale.hidden_states_bias": "model-00001-of-00004.safetensors", + "model.layers.5.res_scale.hidden_states_scale": "model-00001-of-00004.safetensors", + "model.layers.5.res_scale.residual_bias": "model-00001-of-00004.safetensors", + "model.layers.5.res_scale.residual_scale": "model-00001-of-00004.safetensors", + "model.layers.5.zaya_block.experts.local_experts.0.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.5.zaya_block.experts.local_experts.0.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.5.zaya_block.experts.local_experts.1.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.5.zaya_block.experts.local_experts.1.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.5.zaya_block.experts.local_experts.10.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.5.zaya_block.experts.local_experts.10.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.5.zaya_block.experts.local_experts.11.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.5.zaya_block.experts.local_experts.11.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.5.zaya_block.experts.local_experts.12.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.5.zaya_block.experts.local_experts.12.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.5.zaya_block.experts.local_experts.13.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.5.zaya_block.experts.local_experts.13.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.5.zaya_block.experts.local_experts.14.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.5.zaya_block.experts.local_experts.14.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.5.zaya_block.experts.local_experts.15.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.5.zaya_block.experts.local_experts.15.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.5.zaya_block.experts.local_experts.2.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.5.zaya_block.experts.local_experts.2.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.5.zaya_block.experts.local_experts.3.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.5.zaya_block.experts.local_experts.3.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.5.zaya_block.experts.local_experts.4.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.5.zaya_block.experts.local_experts.4.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.5.zaya_block.experts.local_experts.5.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.5.zaya_block.experts.local_experts.5.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.5.zaya_block.experts.local_experts.6.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.5.zaya_block.experts.local_experts.6.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.5.zaya_block.experts.local_experts.7.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.5.zaya_block.experts.local_experts.7.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.5.zaya_block.experts.local_experts.8.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.5.zaya_block.experts.local_experts.8.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.5.zaya_block.experts.local_experts.9.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.5.zaya_block.experts.local_experts.9.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.5.zaya_block.router.balancing_biases": "model-00001-of-00004.safetensors", + "model.layers.5.zaya_block.router.down_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.5.zaya_block.router.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.5.zaya_block.router.rmsnorm_eda.weight": "model-00001-of-00004.safetensors", + "model.layers.5.zaya_block.router.router_mlp.0.bias": "model-00001-of-00004.safetensors", + "model.layers.5.zaya_block.router.router_mlp.0.weight": "model-00001-of-00004.safetensors", + "model.layers.5.zaya_block.router.router_mlp.2.bias": "model-00001-of-00004.safetensors", + "model.layers.5.zaya_block.router.router_mlp.2.weight": "model-00001-of-00004.safetensors", + "model.layers.5.zaya_block.router.router_mlp.4.weight": "model-00001-of-00004.safetensors", + "model.layers.5.zaya_block.router.router_states_scale": "model-00001-of-00004.safetensors", + "model.layers.50.input_norm.weight": "model-00003-of-00004.safetensors", + "model.layers.50.res_scale.hidden_states_bias": "model-00003-of-00004.safetensors", + "model.layers.50.res_scale.hidden_states_scale": "model-00003-of-00004.safetensors", + "model.layers.50.res_scale.residual_bias": "model-00003-of-00004.safetensors", + "model.layers.50.res_scale.residual_scale": "model-00003-of-00004.safetensors", + "model.layers.50.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.50.self_attn.qkv.conv_qk.0.bias": "model-00003-of-00004.safetensors", + "model.layers.50.self_attn.qkv.conv_qk.0.weight": "model-00003-of-00004.safetensors", + "model.layers.50.self_attn.qkv.conv_qk.1.bias": "model-00003-of-00004.safetensors", + "model.layers.50.self_attn.qkv.conv_qk.1.weight": "model-00003-of-00004.safetensors", + "model.layers.50.self_attn.qkv.linear_k.weight": "model-00003-of-00004.safetensors", + "model.layers.50.self_attn.qkv.linear_q.weight": "model-00003-of-00004.safetensors", + "model.layers.50.self_attn.qkv.temp": "model-00003-of-00004.safetensors", + "model.layers.50.self_attn.qkv.val_proj1.weight": "model-00003-of-00004.safetensors", + "model.layers.50.self_attn.qkv.val_proj2.weight": "model-00003-of-00004.safetensors", + "model.layers.51.input_norm.weight": "model-00003-of-00004.safetensors", + "model.layers.51.res_scale.hidden_states_bias": "model-00003-of-00004.safetensors", + "model.layers.51.res_scale.hidden_states_scale": "model-00003-of-00004.safetensors", + "model.layers.51.res_scale.residual_bias": "model-00003-of-00004.safetensors", + "model.layers.51.res_scale.residual_scale": "model-00003-of-00004.safetensors", + "model.layers.51.zaya_block.experts.local_experts.0.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.51.zaya_block.experts.local_experts.0.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.51.zaya_block.experts.local_experts.1.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.51.zaya_block.experts.local_experts.1.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.51.zaya_block.experts.local_experts.10.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.51.zaya_block.experts.local_experts.10.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.51.zaya_block.experts.local_experts.11.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.51.zaya_block.experts.local_experts.11.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.51.zaya_block.experts.local_experts.12.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.51.zaya_block.experts.local_experts.12.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.51.zaya_block.experts.local_experts.13.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.51.zaya_block.experts.local_experts.13.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.51.zaya_block.experts.local_experts.14.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.51.zaya_block.experts.local_experts.14.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.51.zaya_block.experts.local_experts.15.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.51.zaya_block.experts.local_experts.15.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.51.zaya_block.experts.local_experts.2.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.51.zaya_block.experts.local_experts.2.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.51.zaya_block.experts.local_experts.3.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.51.zaya_block.experts.local_experts.3.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.51.zaya_block.experts.local_experts.4.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.51.zaya_block.experts.local_experts.4.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.51.zaya_block.experts.local_experts.5.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.51.zaya_block.experts.local_experts.5.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.51.zaya_block.experts.local_experts.6.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.51.zaya_block.experts.local_experts.6.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.51.zaya_block.experts.local_experts.7.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.51.zaya_block.experts.local_experts.7.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.51.zaya_block.experts.local_experts.8.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.51.zaya_block.experts.local_experts.8.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.51.zaya_block.experts.local_experts.9.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.51.zaya_block.experts.local_experts.9.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.51.zaya_block.router.balancing_biases": "model-00003-of-00004.safetensors", + "model.layers.51.zaya_block.router.down_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.51.zaya_block.router.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.51.zaya_block.router.rmsnorm_eda.weight": "model-00003-of-00004.safetensors", + "model.layers.51.zaya_block.router.router_mlp.0.bias": "model-00003-of-00004.safetensors", + "model.layers.51.zaya_block.router.router_mlp.0.weight": "model-00003-of-00004.safetensors", + "model.layers.51.zaya_block.router.router_mlp.2.bias": "model-00003-of-00004.safetensors", + "model.layers.51.zaya_block.router.router_mlp.2.weight": "model-00003-of-00004.safetensors", + "model.layers.51.zaya_block.router.router_mlp.4.weight": "model-00003-of-00004.safetensors", + "model.layers.51.zaya_block.router.router_states_scale": "model-00003-of-00004.safetensors", + "model.layers.52.input_norm.weight": "model-00003-of-00004.safetensors", + "model.layers.52.res_scale.hidden_states_bias": "model-00003-of-00004.safetensors", + "model.layers.52.res_scale.hidden_states_scale": "model-00003-of-00004.safetensors", + "model.layers.52.res_scale.residual_bias": "model-00003-of-00004.safetensors", + "model.layers.52.res_scale.residual_scale": "model-00003-of-00004.safetensors", + "model.layers.52.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.52.self_attn.qkv.conv_qk.0.bias": "model-00003-of-00004.safetensors", + "model.layers.52.self_attn.qkv.conv_qk.0.weight": "model-00003-of-00004.safetensors", + "model.layers.52.self_attn.qkv.conv_qk.1.bias": "model-00003-of-00004.safetensors", + "model.layers.52.self_attn.qkv.conv_qk.1.weight": "model-00003-of-00004.safetensors", + "model.layers.52.self_attn.qkv.linear_k.weight": "model-00003-of-00004.safetensors", + "model.layers.52.self_attn.qkv.linear_q.weight": "model-00003-of-00004.safetensors", + "model.layers.52.self_attn.qkv.temp": "model-00003-of-00004.safetensors", + "model.layers.52.self_attn.qkv.val_proj1.weight": "model-00003-of-00004.safetensors", + "model.layers.52.self_attn.qkv.val_proj2.weight": "model-00003-of-00004.safetensors", + "model.layers.53.input_norm.weight": "model-00003-of-00004.safetensors", + "model.layers.53.res_scale.hidden_states_bias": "model-00003-of-00004.safetensors", + "model.layers.53.res_scale.hidden_states_scale": "model-00003-of-00004.safetensors", + "model.layers.53.res_scale.residual_bias": "model-00003-of-00004.safetensors", + "model.layers.53.res_scale.residual_scale": "model-00003-of-00004.safetensors", + "model.layers.53.zaya_block.experts.local_experts.0.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.53.zaya_block.experts.local_experts.0.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.53.zaya_block.experts.local_experts.1.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.53.zaya_block.experts.local_experts.1.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.53.zaya_block.experts.local_experts.10.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.53.zaya_block.experts.local_experts.10.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.53.zaya_block.experts.local_experts.11.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.53.zaya_block.experts.local_experts.11.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.53.zaya_block.experts.local_experts.12.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.53.zaya_block.experts.local_experts.12.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.53.zaya_block.experts.local_experts.13.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.53.zaya_block.experts.local_experts.13.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.53.zaya_block.experts.local_experts.14.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.53.zaya_block.experts.local_experts.14.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.53.zaya_block.experts.local_experts.15.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.53.zaya_block.experts.local_experts.15.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.53.zaya_block.experts.local_experts.2.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.53.zaya_block.experts.local_experts.2.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.53.zaya_block.experts.local_experts.3.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.53.zaya_block.experts.local_experts.3.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.53.zaya_block.experts.local_experts.4.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.53.zaya_block.experts.local_experts.4.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.53.zaya_block.experts.local_experts.5.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.53.zaya_block.experts.local_experts.5.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.53.zaya_block.experts.local_experts.6.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.53.zaya_block.experts.local_experts.6.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.53.zaya_block.experts.local_experts.7.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.53.zaya_block.experts.local_experts.7.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.53.zaya_block.experts.local_experts.8.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.53.zaya_block.experts.local_experts.8.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.53.zaya_block.experts.local_experts.9.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.53.zaya_block.experts.local_experts.9.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.53.zaya_block.router.balancing_biases": "model-00003-of-00004.safetensors", + "model.layers.53.zaya_block.router.down_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.53.zaya_block.router.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.53.zaya_block.router.rmsnorm_eda.weight": "model-00003-of-00004.safetensors", + "model.layers.53.zaya_block.router.router_mlp.0.bias": "model-00003-of-00004.safetensors", + "model.layers.53.zaya_block.router.router_mlp.0.weight": "model-00003-of-00004.safetensors", + "model.layers.53.zaya_block.router.router_mlp.2.bias": "model-00003-of-00004.safetensors", + "model.layers.53.zaya_block.router.router_mlp.2.weight": "model-00003-of-00004.safetensors", + "model.layers.53.zaya_block.router.router_mlp.4.weight": "model-00003-of-00004.safetensors", + "model.layers.53.zaya_block.router.router_states_scale": "model-00003-of-00004.safetensors", + "model.layers.54.input_norm.weight": "model-00003-of-00004.safetensors", + "model.layers.54.res_scale.hidden_states_bias": "model-00003-of-00004.safetensors", + "model.layers.54.res_scale.hidden_states_scale": "model-00003-of-00004.safetensors", + "model.layers.54.res_scale.residual_bias": "model-00003-of-00004.safetensors", + "model.layers.54.res_scale.residual_scale": "model-00003-of-00004.safetensors", + "model.layers.54.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.54.self_attn.qkv.conv_qk.0.bias": "model-00003-of-00004.safetensors", + "model.layers.54.self_attn.qkv.conv_qk.0.weight": "model-00003-of-00004.safetensors", + "model.layers.54.self_attn.qkv.conv_qk.1.bias": "model-00003-of-00004.safetensors", + "model.layers.54.self_attn.qkv.conv_qk.1.weight": "model-00003-of-00004.safetensors", + "model.layers.54.self_attn.qkv.linear_k.weight": "model-00003-of-00004.safetensors", + "model.layers.54.self_attn.qkv.linear_q.weight": "model-00003-of-00004.safetensors", + "model.layers.54.self_attn.qkv.temp": "model-00003-of-00004.safetensors", + "model.layers.54.self_attn.qkv.val_proj1.weight": "model-00003-of-00004.safetensors", + "model.layers.54.self_attn.qkv.val_proj2.weight": "model-00003-of-00004.safetensors", + "model.layers.55.input_norm.weight": "model-00003-of-00004.safetensors", + "model.layers.55.res_scale.hidden_states_bias": "model-00003-of-00004.safetensors", + "model.layers.55.res_scale.hidden_states_scale": "model-00003-of-00004.safetensors", + "model.layers.55.res_scale.residual_bias": "model-00003-of-00004.safetensors", + "model.layers.55.res_scale.residual_scale": "model-00003-of-00004.safetensors", + "model.layers.55.zaya_block.experts.local_experts.0.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.55.zaya_block.experts.local_experts.0.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.55.zaya_block.experts.local_experts.1.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.55.zaya_block.experts.local_experts.1.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.55.zaya_block.experts.local_experts.10.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.55.zaya_block.experts.local_experts.10.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.55.zaya_block.experts.local_experts.11.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.55.zaya_block.experts.local_experts.11.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.55.zaya_block.experts.local_experts.12.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.55.zaya_block.experts.local_experts.12.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.55.zaya_block.experts.local_experts.13.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.55.zaya_block.experts.local_experts.13.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.55.zaya_block.experts.local_experts.14.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.55.zaya_block.experts.local_experts.14.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.55.zaya_block.experts.local_experts.15.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.55.zaya_block.experts.local_experts.15.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.55.zaya_block.experts.local_experts.2.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.55.zaya_block.experts.local_experts.2.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.55.zaya_block.experts.local_experts.3.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.55.zaya_block.experts.local_experts.3.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.55.zaya_block.experts.local_experts.4.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.55.zaya_block.experts.local_experts.4.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.55.zaya_block.experts.local_experts.5.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.55.zaya_block.experts.local_experts.5.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.55.zaya_block.experts.local_experts.6.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.55.zaya_block.experts.local_experts.6.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.55.zaya_block.experts.local_experts.7.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.55.zaya_block.experts.local_experts.7.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.55.zaya_block.experts.local_experts.8.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.55.zaya_block.experts.local_experts.8.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.55.zaya_block.experts.local_experts.9.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.55.zaya_block.experts.local_experts.9.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.55.zaya_block.router.balancing_biases": "model-00003-of-00004.safetensors", + "model.layers.55.zaya_block.router.down_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.55.zaya_block.router.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.55.zaya_block.router.rmsnorm_eda.weight": "model-00003-of-00004.safetensors", + "model.layers.55.zaya_block.router.router_mlp.0.bias": "model-00003-of-00004.safetensors", + "model.layers.55.zaya_block.router.router_mlp.0.weight": "model-00003-of-00004.safetensors", + "model.layers.55.zaya_block.router.router_mlp.2.bias": "model-00003-of-00004.safetensors", + "model.layers.55.zaya_block.router.router_mlp.2.weight": "model-00003-of-00004.safetensors", + "model.layers.55.zaya_block.router.router_mlp.4.weight": "model-00003-of-00004.safetensors", + "model.layers.55.zaya_block.router.router_states_scale": "model-00003-of-00004.safetensors", + "model.layers.56.input_norm.weight": "model-00003-of-00004.safetensors", + "model.layers.56.res_scale.hidden_states_bias": "model-00003-of-00004.safetensors", + "model.layers.56.res_scale.hidden_states_scale": "model-00003-of-00004.safetensors", + "model.layers.56.res_scale.residual_bias": "model-00003-of-00004.safetensors", + "model.layers.56.res_scale.residual_scale": "model-00003-of-00004.safetensors", + "model.layers.56.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.56.self_attn.qkv.conv_qk.0.bias": "model-00003-of-00004.safetensors", + "model.layers.56.self_attn.qkv.conv_qk.0.weight": "model-00003-of-00004.safetensors", + "model.layers.56.self_attn.qkv.conv_qk.1.bias": "model-00003-of-00004.safetensors", + "model.layers.56.self_attn.qkv.conv_qk.1.weight": "model-00003-of-00004.safetensors", + "model.layers.56.self_attn.qkv.linear_k.weight": "model-00003-of-00004.safetensors", + "model.layers.56.self_attn.qkv.linear_q.weight": "model-00003-of-00004.safetensors", + "model.layers.56.self_attn.qkv.temp": "model-00003-of-00004.safetensors", + "model.layers.56.self_attn.qkv.val_proj1.weight": "model-00003-of-00004.safetensors", + "model.layers.56.self_attn.qkv.val_proj2.weight": "model-00003-of-00004.safetensors", + "model.layers.57.input_norm.weight": "model-00003-of-00004.safetensors", + "model.layers.57.res_scale.hidden_states_bias": "model-00003-of-00004.safetensors", + "model.layers.57.res_scale.hidden_states_scale": "model-00003-of-00004.safetensors", + "model.layers.57.res_scale.residual_bias": "model-00003-of-00004.safetensors", + "model.layers.57.res_scale.residual_scale": "model-00003-of-00004.safetensors", + "model.layers.57.zaya_block.experts.local_experts.0.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.57.zaya_block.experts.local_experts.0.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.57.zaya_block.experts.local_experts.1.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.57.zaya_block.experts.local_experts.1.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.57.zaya_block.experts.local_experts.10.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.57.zaya_block.experts.local_experts.10.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.57.zaya_block.experts.local_experts.11.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.57.zaya_block.experts.local_experts.11.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.57.zaya_block.experts.local_experts.12.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.57.zaya_block.experts.local_experts.12.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.57.zaya_block.experts.local_experts.13.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.57.zaya_block.experts.local_experts.13.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.57.zaya_block.experts.local_experts.14.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.57.zaya_block.experts.local_experts.14.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.57.zaya_block.experts.local_experts.15.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.57.zaya_block.experts.local_experts.15.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.57.zaya_block.experts.local_experts.2.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.57.zaya_block.experts.local_experts.2.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.57.zaya_block.experts.local_experts.3.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.57.zaya_block.experts.local_experts.3.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.57.zaya_block.experts.local_experts.4.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.57.zaya_block.experts.local_experts.4.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.57.zaya_block.experts.local_experts.5.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.57.zaya_block.experts.local_experts.5.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.57.zaya_block.experts.local_experts.6.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.57.zaya_block.experts.local_experts.6.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.57.zaya_block.experts.local_experts.7.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.57.zaya_block.experts.local_experts.7.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.57.zaya_block.experts.local_experts.8.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.57.zaya_block.experts.local_experts.8.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.57.zaya_block.experts.local_experts.9.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.57.zaya_block.experts.local_experts.9.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.57.zaya_block.router.balancing_biases": "model-00003-of-00004.safetensors", + "model.layers.57.zaya_block.router.down_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.57.zaya_block.router.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.57.zaya_block.router.rmsnorm_eda.weight": "model-00003-of-00004.safetensors", + "model.layers.57.zaya_block.router.router_mlp.0.bias": "model-00003-of-00004.safetensors", + "model.layers.57.zaya_block.router.router_mlp.0.weight": "model-00003-of-00004.safetensors", + "model.layers.57.zaya_block.router.router_mlp.2.bias": "model-00003-of-00004.safetensors", + "model.layers.57.zaya_block.router.router_mlp.2.weight": "model-00003-of-00004.safetensors", + "model.layers.57.zaya_block.router.router_mlp.4.weight": "model-00003-of-00004.safetensors", + "model.layers.57.zaya_block.router.router_states_scale": "model-00003-of-00004.safetensors", + "model.layers.58.input_norm.weight": "model-00003-of-00004.safetensors", + "model.layers.58.res_scale.hidden_states_bias": "model-00003-of-00004.safetensors", + "model.layers.58.res_scale.hidden_states_scale": "model-00003-of-00004.safetensors", + "model.layers.58.res_scale.residual_bias": "model-00003-of-00004.safetensors", + "model.layers.58.res_scale.residual_scale": "model-00003-of-00004.safetensors", + "model.layers.58.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.58.self_attn.qkv.conv_qk.0.bias": "model-00003-of-00004.safetensors", + "model.layers.58.self_attn.qkv.conv_qk.0.weight": "model-00003-of-00004.safetensors", + "model.layers.58.self_attn.qkv.conv_qk.1.bias": "model-00003-of-00004.safetensors", + "model.layers.58.self_attn.qkv.conv_qk.1.weight": "model-00003-of-00004.safetensors", + "model.layers.58.self_attn.qkv.linear_k.weight": "model-00003-of-00004.safetensors", + "model.layers.58.self_attn.qkv.linear_q.weight": "model-00003-of-00004.safetensors", + "model.layers.58.self_attn.qkv.temp": "model-00003-of-00004.safetensors", + "model.layers.58.self_attn.qkv.val_proj1.weight": "model-00003-of-00004.safetensors", + "model.layers.58.self_attn.qkv.val_proj2.weight": "model-00003-of-00004.safetensors", + "model.layers.59.input_norm.weight": "model-00003-of-00004.safetensors", + "model.layers.59.res_scale.hidden_states_bias": "model-00003-of-00004.safetensors", + "model.layers.59.res_scale.hidden_states_scale": "model-00003-of-00004.safetensors", + "model.layers.59.res_scale.residual_bias": "model-00003-of-00004.safetensors", + "model.layers.59.res_scale.residual_scale": "model-00003-of-00004.safetensors", + "model.layers.59.zaya_block.experts.local_experts.0.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.59.zaya_block.experts.local_experts.0.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.59.zaya_block.experts.local_experts.1.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.59.zaya_block.experts.local_experts.1.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.59.zaya_block.experts.local_experts.10.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.59.zaya_block.experts.local_experts.10.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.59.zaya_block.experts.local_experts.11.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.59.zaya_block.experts.local_experts.11.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.59.zaya_block.experts.local_experts.12.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.59.zaya_block.experts.local_experts.12.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.59.zaya_block.experts.local_experts.13.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.59.zaya_block.experts.local_experts.13.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.59.zaya_block.experts.local_experts.14.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.59.zaya_block.experts.local_experts.14.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.59.zaya_block.experts.local_experts.15.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.59.zaya_block.experts.local_experts.15.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.59.zaya_block.experts.local_experts.2.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.59.zaya_block.experts.local_experts.2.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.59.zaya_block.experts.local_experts.3.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.59.zaya_block.experts.local_experts.3.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.59.zaya_block.experts.local_experts.4.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.59.zaya_block.experts.local_experts.4.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.59.zaya_block.experts.local_experts.5.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.59.zaya_block.experts.local_experts.5.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.59.zaya_block.experts.local_experts.6.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.59.zaya_block.experts.local_experts.6.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.59.zaya_block.experts.local_experts.7.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.59.zaya_block.experts.local_experts.7.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.59.zaya_block.experts.local_experts.8.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.59.zaya_block.experts.local_experts.8.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.59.zaya_block.experts.local_experts.9.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.59.zaya_block.experts.local_experts.9.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.59.zaya_block.router.balancing_biases": "model-00003-of-00004.safetensors", + "model.layers.59.zaya_block.router.down_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.59.zaya_block.router.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.59.zaya_block.router.rmsnorm_eda.weight": "model-00003-of-00004.safetensors", + "model.layers.59.zaya_block.router.router_mlp.0.bias": "model-00003-of-00004.safetensors", + "model.layers.59.zaya_block.router.router_mlp.0.weight": "model-00003-of-00004.safetensors", + "model.layers.59.zaya_block.router.router_mlp.2.bias": "model-00003-of-00004.safetensors", + "model.layers.59.zaya_block.router.router_mlp.2.weight": "model-00003-of-00004.safetensors", + "model.layers.59.zaya_block.router.router_mlp.4.weight": "model-00003-of-00004.safetensors", + "model.layers.59.zaya_block.router.router_states_scale": "model-00003-of-00004.safetensors", + "model.layers.6.input_norm.weight": "model-00001-of-00004.safetensors", + "model.layers.6.res_scale.hidden_states_bias": "model-00001-of-00004.safetensors", + "model.layers.6.res_scale.hidden_states_scale": "model-00001-of-00004.safetensors", + "model.layers.6.res_scale.residual_bias": "model-00001-of-00004.safetensors", + "model.layers.6.res_scale.residual_scale": "model-00001-of-00004.safetensors", + "model.layers.6.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.6.self_attn.qkv.conv_qk.0.bias": "model-00001-of-00004.safetensors", + "model.layers.6.self_attn.qkv.conv_qk.0.weight": "model-00001-of-00004.safetensors", + "model.layers.6.self_attn.qkv.conv_qk.1.bias": "model-00001-of-00004.safetensors", + "model.layers.6.self_attn.qkv.conv_qk.1.weight": "model-00001-of-00004.safetensors", + "model.layers.6.self_attn.qkv.linear_k.weight": "model-00001-of-00004.safetensors", + "model.layers.6.self_attn.qkv.linear_q.weight": "model-00001-of-00004.safetensors", + "model.layers.6.self_attn.qkv.temp": "model-00001-of-00004.safetensors", + "model.layers.6.self_attn.qkv.val_proj1.weight": "model-00001-of-00004.safetensors", + "model.layers.6.self_attn.qkv.val_proj2.weight": "model-00001-of-00004.safetensors", + "model.layers.60.input_norm.weight": "model-00003-of-00004.safetensors", + "model.layers.60.res_scale.hidden_states_bias": "model-00003-of-00004.safetensors", + "model.layers.60.res_scale.hidden_states_scale": "model-00003-of-00004.safetensors", + "model.layers.60.res_scale.residual_bias": "model-00003-of-00004.safetensors", + "model.layers.60.res_scale.residual_scale": "model-00003-of-00004.safetensors", + "model.layers.60.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.60.self_attn.qkv.conv_qk.0.bias": "model-00003-of-00004.safetensors", + "model.layers.60.self_attn.qkv.conv_qk.0.weight": "model-00003-of-00004.safetensors", + "model.layers.60.self_attn.qkv.conv_qk.1.bias": "model-00003-of-00004.safetensors", + "model.layers.60.self_attn.qkv.conv_qk.1.weight": "model-00003-of-00004.safetensors", + "model.layers.60.self_attn.qkv.linear_k.weight": "model-00003-of-00004.safetensors", + "model.layers.60.self_attn.qkv.linear_q.weight": "model-00003-of-00004.safetensors", + "model.layers.60.self_attn.qkv.temp": "model-00003-of-00004.safetensors", + "model.layers.60.self_attn.qkv.val_proj1.weight": "model-00003-of-00004.safetensors", + "model.layers.60.self_attn.qkv.val_proj2.weight": "model-00003-of-00004.safetensors", + "model.layers.61.input_norm.weight": "model-00003-of-00004.safetensors", + "model.layers.61.res_scale.hidden_states_bias": "model-00003-of-00004.safetensors", + "model.layers.61.res_scale.hidden_states_scale": "model-00003-of-00004.safetensors", + "model.layers.61.res_scale.residual_bias": "model-00003-of-00004.safetensors", + "model.layers.61.res_scale.residual_scale": "model-00003-of-00004.safetensors", + "model.layers.61.zaya_block.experts.local_experts.0.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.61.zaya_block.experts.local_experts.0.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.61.zaya_block.experts.local_experts.1.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.61.zaya_block.experts.local_experts.1.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.61.zaya_block.experts.local_experts.10.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.61.zaya_block.experts.local_experts.10.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.61.zaya_block.experts.local_experts.11.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.61.zaya_block.experts.local_experts.11.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.61.zaya_block.experts.local_experts.12.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.61.zaya_block.experts.local_experts.12.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.61.zaya_block.experts.local_experts.13.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.61.zaya_block.experts.local_experts.13.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.61.zaya_block.experts.local_experts.14.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.61.zaya_block.experts.local_experts.14.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.61.zaya_block.experts.local_experts.15.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.61.zaya_block.experts.local_experts.15.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.61.zaya_block.experts.local_experts.2.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.61.zaya_block.experts.local_experts.2.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.61.zaya_block.experts.local_experts.3.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.61.zaya_block.experts.local_experts.3.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.61.zaya_block.experts.local_experts.4.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.61.zaya_block.experts.local_experts.4.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.61.zaya_block.experts.local_experts.5.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.61.zaya_block.experts.local_experts.5.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.61.zaya_block.experts.local_experts.6.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.61.zaya_block.experts.local_experts.6.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.61.zaya_block.experts.local_experts.7.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.61.zaya_block.experts.local_experts.7.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.61.zaya_block.experts.local_experts.8.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.61.zaya_block.experts.local_experts.8.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.61.zaya_block.experts.local_experts.9.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.61.zaya_block.experts.local_experts.9.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.61.zaya_block.router.balancing_biases": "model-00003-of-00004.safetensors", + "model.layers.61.zaya_block.router.down_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.61.zaya_block.router.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.61.zaya_block.router.rmsnorm_eda.weight": "model-00003-of-00004.safetensors", + "model.layers.61.zaya_block.router.router_mlp.0.bias": "model-00003-of-00004.safetensors", + "model.layers.61.zaya_block.router.router_mlp.0.weight": "model-00003-of-00004.safetensors", + "model.layers.61.zaya_block.router.router_mlp.2.bias": "model-00003-of-00004.safetensors", + "model.layers.61.zaya_block.router.router_mlp.2.weight": "model-00003-of-00004.safetensors", + "model.layers.61.zaya_block.router.router_mlp.4.weight": "model-00003-of-00004.safetensors", + "model.layers.61.zaya_block.router.router_states_scale": "model-00003-of-00004.safetensors", + "model.layers.62.input_norm.weight": "model-00003-of-00004.safetensors", + "model.layers.62.res_scale.hidden_states_bias": "model-00003-of-00004.safetensors", + "model.layers.62.res_scale.hidden_states_scale": "model-00003-of-00004.safetensors", + "model.layers.62.res_scale.residual_bias": "model-00003-of-00004.safetensors", + "model.layers.62.res_scale.residual_scale": "model-00003-of-00004.safetensors", + "model.layers.62.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.62.self_attn.qkv.conv_qk.0.bias": "model-00003-of-00004.safetensors", + "model.layers.62.self_attn.qkv.conv_qk.0.weight": "model-00003-of-00004.safetensors", + "model.layers.62.self_attn.qkv.conv_qk.1.bias": "model-00003-of-00004.safetensors", + "model.layers.62.self_attn.qkv.conv_qk.1.weight": "model-00003-of-00004.safetensors", + "model.layers.62.self_attn.qkv.linear_k.weight": "model-00003-of-00004.safetensors", + "model.layers.62.self_attn.qkv.linear_q.weight": "model-00003-of-00004.safetensors", + "model.layers.62.self_attn.qkv.temp": "model-00003-of-00004.safetensors", + "model.layers.62.self_attn.qkv.val_proj1.weight": "model-00003-of-00004.safetensors", + "model.layers.62.self_attn.qkv.val_proj2.weight": "model-00003-of-00004.safetensors", + "model.layers.63.input_norm.weight": "model-00003-of-00004.safetensors", + "model.layers.63.res_scale.hidden_states_bias": "model-00003-of-00004.safetensors", + "model.layers.63.res_scale.hidden_states_scale": "model-00003-of-00004.safetensors", + "model.layers.63.res_scale.residual_bias": "model-00003-of-00004.safetensors", + "model.layers.63.res_scale.residual_scale": "model-00003-of-00004.safetensors", + "model.layers.63.zaya_block.experts.local_experts.0.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.63.zaya_block.experts.local_experts.0.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.63.zaya_block.experts.local_experts.1.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.63.zaya_block.experts.local_experts.1.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.63.zaya_block.experts.local_experts.10.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.63.zaya_block.experts.local_experts.10.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.63.zaya_block.experts.local_experts.11.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.63.zaya_block.experts.local_experts.11.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.63.zaya_block.experts.local_experts.12.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.63.zaya_block.experts.local_experts.12.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.63.zaya_block.experts.local_experts.13.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.63.zaya_block.experts.local_experts.13.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.63.zaya_block.experts.local_experts.14.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.63.zaya_block.experts.local_experts.14.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.63.zaya_block.experts.local_experts.15.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.63.zaya_block.experts.local_experts.15.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.63.zaya_block.experts.local_experts.2.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.63.zaya_block.experts.local_experts.2.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.63.zaya_block.experts.local_experts.3.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.63.zaya_block.experts.local_experts.3.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.63.zaya_block.experts.local_experts.4.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.63.zaya_block.experts.local_experts.4.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.63.zaya_block.experts.local_experts.5.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.63.zaya_block.experts.local_experts.5.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.63.zaya_block.experts.local_experts.6.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.63.zaya_block.experts.local_experts.6.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.63.zaya_block.experts.local_experts.7.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.63.zaya_block.experts.local_experts.7.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.63.zaya_block.experts.local_experts.8.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.63.zaya_block.experts.local_experts.8.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.63.zaya_block.experts.local_experts.9.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.63.zaya_block.experts.local_experts.9.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.63.zaya_block.router.balancing_biases": "model-00003-of-00004.safetensors", + "model.layers.63.zaya_block.router.down_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.63.zaya_block.router.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.63.zaya_block.router.rmsnorm_eda.weight": "model-00003-of-00004.safetensors", + "model.layers.63.zaya_block.router.router_mlp.0.bias": "model-00003-of-00004.safetensors", + "model.layers.63.zaya_block.router.router_mlp.0.weight": "model-00003-of-00004.safetensors", + "model.layers.63.zaya_block.router.router_mlp.2.bias": "model-00003-of-00004.safetensors", + "model.layers.63.zaya_block.router.router_mlp.2.weight": "model-00003-of-00004.safetensors", + "model.layers.63.zaya_block.router.router_mlp.4.weight": "model-00003-of-00004.safetensors", + "model.layers.63.zaya_block.router.router_states_scale": "model-00003-of-00004.safetensors", + "model.layers.64.input_norm.weight": "model-00003-of-00004.safetensors", + "model.layers.64.res_scale.hidden_states_bias": "model-00003-of-00004.safetensors", + "model.layers.64.res_scale.hidden_states_scale": "model-00003-of-00004.safetensors", + "model.layers.64.res_scale.residual_bias": "model-00003-of-00004.safetensors", + "model.layers.64.res_scale.residual_scale": "model-00003-of-00004.safetensors", + "model.layers.64.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.64.self_attn.qkv.conv_qk.0.bias": "model-00003-of-00004.safetensors", + "model.layers.64.self_attn.qkv.conv_qk.0.weight": "model-00003-of-00004.safetensors", + "model.layers.64.self_attn.qkv.conv_qk.1.bias": "model-00003-of-00004.safetensors", + "model.layers.64.self_attn.qkv.conv_qk.1.weight": "model-00003-of-00004.safetensors", + "model.layers.64.self_attn.qkv.linear_k.weight": "model-00003-of-00004.safetensors", + "model.layers.64.self_attn.qkv.linear_q.weight": "model-00003-of-00004.safetensors", + "model.layers.64.self_attn.qkv.temp": "model-00003-of-00004.safetensors", + "model.layers.64.self_attn.qkv.val_proj1.weight": "model-00003-of-00004.safetensors", + "model.layers.64.self_attn.qkv.val_proj2.weight": "model-00003-of-00004.safetensors", + "model.layers.65.input_norm.weight": "model-00003-of-00004.safetensors", + "model.layers.65.res_scale.hidden_states_bias": "model-00003-of-00004.safetensors", + "model.layers.65.res_scale.hidden_states_scale": "model-00003-of-00004.safetensors", + "model.layers.65.res_scale.residual_bias": "model-00003-of-00004.safetensors", + "model.layers.65.res_scale.residual_scale": "model-00003-of-00004.safetensors", + "model.layers.65.zaya_block.experts.local_experts.0.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.65.zaya_block.experts.local_experts.0.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.65.zaya_block.experts.local_experts.1.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.65.zaya_block.experts.local_experts.1.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.65.zaya_block.experts.local_experts.10.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.65.zaya_block.experts.local_experts.10.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.65.zaya_block.experts.local_experts.11.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.65.zaya_block.experts.local_experts.11.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.65.zaya_block.experts.local_experts.12.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.65.zaya_block.experts.local_experts.12.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.65.zaya_block.experts.local_experts.13.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.65.zaya_block.experts.local_experts.13.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.65.zaya_block.experts.local_experts.14.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.65.zaya_block.experts.local_experts.14.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.65.zaya_block.experts.local_experts.15.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.65.zaya_block.experts.local_experts.15.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.65.zaya_block.experts.local_experts.2.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.65.zaya_block.experts.local_experts.2.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.65.zaya_block.experts.local_experts.3.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.65.zaya_block.experts.local_experts.3.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.65.zaya_block.experts.local_experts.4.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.65.zaya_block.experts.local_experts.4.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.65.zaya_block.experts.local_experts.5.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.65.zaya_block.experts.local_experts.5.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.65.zaya_block.experts.local_experts.6.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.65.zaya_block.experts.local_experts.6.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.65.zaya_block.experts.local_experts.7.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.65.zaya_block.experts.local_experts.7.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.65.zaya_block.experts.local_experts.8.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.65.zaya_block.experts.local_experts.8.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.65.zaya_block.experts.local_experts.9.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.65.zaya_block.experts.local_experts.9.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.65.zaya_block.router.balancing_biases": "model-00003-of-00004.safetensors", + "model.layers.65.zaya_block.router.down_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.65.zaya_block.router.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.65.zaya_block.router.rmsnorm_eda.weight": "model-00003-of-00004.safetensors", + "model.layers.65.zaya_block.router.router_mlp.0.bias": "model-00003-of-00004.safetensors", + "model.layers.65.zaya_block.router.router_mlp.0.weight": "model-00003-of-00004.safetensors", + "model.layers.65.zaya_block.router.router_mlp.2.bias": "model-00003-of-00004.safetensors", + "model.layers.65.zaya_block.router.router_mlp.2.weight": "model-00003-of-00004.safetensors", + "model.layers.65.zaya_block.router.router_mlp.4.weight": "model-00003-of-00004.safetensors", + "model.layers.65.zaya_block.router.router_states_scale": "model-00003-of-00004.safetensors", + "model.layers.66.input_norm.weight": "model-00003-of-00004.safetensors", + "model.layers.66.res_scale.hidden_states_bias": "model-00003-of-00004.safetensors", + "model.layers.66.res_scale.hidden_states_scale": "model-00003-of-00004.safetensors", + "model.layers.66.res_scale.residual_bias": "model-00003-of-00004.safetensors", + "model.layers.66.res_scale.residual_scale": "model-00003-of-00004.safetensors", + "model.layers.66.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.66.self_attn.qkv.conv_qk.0.bias": "model-00003-of-00004.safetensors", + "model.layers.66.self_attn.qkv.conv_qk.0.weight": "model-00003-of-00004.safetensors", + "model.layers.66.self_attn.qkv.conv_qk.1.bias": "model-00003-of-00004.safetensors", + "model.layers.66.self_attn.qkv.conv_qk.1.weight": "model-00003-of-00004.safetensors", + "model.layers.66.self_attn.qkv.linear_k.weight": "model-00003-of-00004.safetensors", + "model.layers.66.self_attn.qkv.linear_q.weight": "model-00003-of-00004.safetensors", + "model.layers.66.self_attn.qkv.temp": "model-00003-of-00004.safetensors", + "model.layers.66.self_attn.qkv.val_proj1.weight": "model-00003-of-00004.safetensors", + "model.layers.66.self_attn.qkv.val_proj2.weight": "model-00003-of-00004.safetensors", + "model.layers.67.input_norm.weight": "model-00004-of-00004.safetensors", + "model.layers.67.res_scale.hidden_states_bias": "model-00004-of-00004.safetensors", + "model.layers.67.res_scale.hidden_states_scale": "model-00004-of-00004.safetensors", + "model.layers.67.res_scale.residual_bias": "model-00004-of-00004.safetensors", + "model.layers.67.res_scale.residual_scale": "model-00004-of-00004.safetensors", + "model.layers.67.zaya_block.experts.local_experts.0.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.67.zaya_block.experts.local_experts.0.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.67.zaya_block.experts.local_experts.1.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.67.zaya_block.experts.local_experts.1.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.67.zaya_block.experts.local_experts.10.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.67.zaya_block.experts.local_experts.10.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.67.zaya_block.experts.local_experts.11.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.67.zaya_block.experts.local_experts.11.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.67.zaya_block.experts.local_experts.12.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.67.zaya_block.experts.local_experts.12.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.67.zaya_block.experts.local_experts.13.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.67.zaya_block.experts.local_experts.13.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.67.zaya_block.experts.local_experts.14.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.67.zaya_block.experts.local_experts.14.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.67.zaya_block.experts.local_experts.15.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.67.zaya_block.experts.local_experts.15.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.67.zaya_block.experts.local_experts.2.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.67.zaya_block.experts.local_experts.2.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.67.zaya_block.experts.local_experts.3.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.67.zaya_block.experts.local_experts.3.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.67.zaya_block.experts.local_experts.4.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.67.zaya_block.experts.local_experts.4.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.67.zaya_block.experts.local_experts.5.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.67.zaya_block.experts.local_experts.5.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.67.zaya_block.experts.local_experts.6.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.67.zaya_block.experts.local_experts.6.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.67.zaya_block.experts.local_experts.7.linear_fc1.weight": "model-00003-of-00004.safetensors", + "model.layers.67.zaya_block.experts.local_experts.7.linear_fc2.weight": "model-00003-of-00004.safetensors", + "model.layers.67.zaya_block.experts.local_experts.8.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.67.zaya_block.experts.local_experts.8.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.67.zaya_block.experts.local_experts.9.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.67.zaya_block.experts.local_experts.9.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.67.zaya_block.router.balancing_biases": "model-00003-of-00004.safetensors", + "model.layers.67.zaya_block.router.down_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.67.zaya_block.router.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.67.zaya_block.router.rmsnorm_eda.weight": "model-00003-of-00004.safetensors", + "model.layers.67.zaya_block.router.router_mlp.0.bias": "model-00003-of-00004.safetensors", + "model.layers.67.zaya_block.router.router_mlp.0.weight": "model-00003-of-00004.safetensors", + "model.layers.67.zaya_block.router.router_mlp.2.bias": "model-00003-of-00004.safetensors", + "model.layers.67.zaya_block.router.router_mlp.2.weight": "model-00003-of-00004.safetensors", + "model.layers.67.zaya_block.router.router_mlp.4.weight": "model-00003-of-00004.safetensors", + "model.layers.67.zaya_block.router.router_states_scale": "model-00003-of-00004.safetensors", + "model.layers.68.input_norm.weight": "model-00004-of-00004.safetensors", + "model.layers.68.res_scale.hidden_states_bias": "model-00004-of-00004.safetensors", + "model.layers.68.res_scale.hidden_states_scale": "model-00004-of-00004.safetensors", + "model.layers.68.res_scale.residual_bias": "model-00004-of-00004.safetensors", + "model.layers.68.res_scale.residual_scale": "model-00004-of-00004.safetensors", + "model.layers.68.self_attn.o_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.68.self_attn.qkv.conv_qk.0.bias": "model-00004-of-00004.safetensors", + "model.layers.68.self_attn.qkv.conv_qk.0.weight": "model-00004-of-00004.safetensors", + "model.layers.68.self_attn.qkv.conv_qk.1.bias": "model-00004-of-00004.safetensors", + "model.layers.68.self_attn.qkv.conv_qk.1.weight": "model-00004-of-00004.safetensors", + "model.layers.68.self_attn.qkv.linear_k.weight": "model-00004-of-00004.safetensors", + "model.layers.68.self_attn.qkv.linear_q.weight": "model-00004-of-00004.safetensors", + "model.layers.68.self_attn.qkv.temp": "model-00004-of-00004.safetensors", + "model.layers.68.self_attn.qkv.val_proj1.weight": "model-00004-of-00004.safetensors", + "model.layers.68.self_attn.qkv.val_proj2.weight": "model-00004-of-00004.safetensors", + "model.layers.69.input_norm.weight": "model-00004-of-00004.safetensors", + "model.layers.69.res_scale.hidden_states_bias": "model-00004-of-00004.safetensors", + "model.layers.69.res_scale.hidden_states_scale": "model-00004-of-00004.safetensors", + "model.layers.69.res_scale.residual_bias": "model-00004-of-00004.safetensors", + "model.layers.69.res_scale.residual_scale": "model-00004-of-00004.safetensors", + "model.layers.69.zaya_block.experts.local_experts.0.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.69.zaya_block.experts.local_experts.0.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.69.zaya_block.experts.local_experts.1.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.69.zaya_block.experts.local_experts.1.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.69.zaya_block.experts.local_experts.10.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.69.zaya_block.experts.local_experts.10.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.69.zaya_block.experts.local_experts.11.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.69.zaya_block.experts.local_experts.11.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.69.zaya_block.experts.local_experts.12.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.69.zaya_block.experts.local_experts.12.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.69.zaya_block.experts.local_experts.13.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.69.zaya_block.experts.local_experts.13.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.69.zaya_block.experts.local_experts.14.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.69.zaya_block.experts.local_experts.14.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.69.zaya_block.experts.local_experts.15.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.69.zaya_block.experts.local_experts.15.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.69.zaya_block.experts.local_experts.2.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.69.zaya_block.experts.local_experts.2.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.69.zaya_block.experts.local_experts.3.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.69.zaya_block.experts.local_experts.3.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.69.zaya_block.experts.local_experts.4.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.69.zaya_block.experts.local_experts.4.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.69.zaya_block.experts.local_experts.5.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.69.zaya_block.experts.local_experts.5.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.69.zaya_block.experts.local_experts.6.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.69.zaya_block.experts.local_experts.6.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.69.zaya_block.experts.local_experts.7.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.69.zaya_block.experts.local_experts.7.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.69.zaya_block.experts.local_experts.8.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.69.zaya_block.experts.local_experts.8.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.69.zaya_block.experts.local_experts.9.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.69.zaya_block.experts.local_experts.9.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.69.zaya_block.router.balancing_biases": "model-00004-of-00004.safetensors", + "model.layers.69.zaya_block.router.down_proj.bias": "model-00004-of-00004.safetensors", + "model.layers.69.zaya_block.router.down_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.69.zaya_block.router.rmsnorm_eda.weight": "model-00004-of-00004.safetensors", + "model.layers.69.zaya_block.router.router_mlp.0.bias": "model-00004-of-00004.safetensors", + "model.layers.69.zaya_block.router.router_mlp.0.weight": "model-00004-of-00004.safetensors", + "model.layers.69.zaya_block.router.router_mlp.2.bias": "model-00004-of-00004.safetensors", + "model.layers.69.zaya_block.router.router_mlp.2.weight": "model-00004-of-00004.safetensors", + "model.layers.69.zaya_block.router.router_mlp.4.weight": "model-00004-of-00004.safetensors", + "model.layers.69.zaya_block.router.router_states_scale": "model-00004-of-00004.safetensors", + "model.layers.7.input_norm.weight": "model-00001-of-00004.safetensors", + "model.layers.7.res_scale.hidden_states_bias": "model-00001-of-00004.safetensors", + "model.layers.7.res_scale.hidden_states_scale": "model-00001-of-00004.safetensors", + "model.layers.7.res_scale.residual_bias": "model-00001-of-00004.safetensors", + "model.layers.7.res_scale.residual_scale": "model-00001-of-00004.safetensors", + "model.layers.7.zaya_block.experts.local_experts.0.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.7.zaya_block.experts.local_experts.0.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.7.zaya_block.experts.local_experts.1.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.7.zaya_block.experts.local_experts.1.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.7.zaya_block.experts.local_experts.10.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.7.zaya_block.experts.local_experts.10.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.7.zaya_block.experts.local_experts.11.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.7.zaya_block.experts.local_experts.11.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.7.zaya_block.experts.local_experts.12.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.7.zaya_block.experts.local_experts.12.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.7.zaya_block.experts.local_experts.13.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.7.zaya_block.experts.local_experts.13.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.7.zaya_block.experts.local_experts.14.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.7.zaya_block.experts.local_experts.14.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.7.zaya_block.experts.local_experts.15.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.7.zaya_block.experts.local_experts.15.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.7.zaya_block.experts.local_experts.2.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.7.zaya_block.experts.local_experts.2.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.7.zaya_block.experts.local_experts.3.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.7.zaya_block.experts.local_experts.3.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.7.zaya_block.experts.local_experts.4.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.7.zaya_block.experts.local_experts.4.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.7.zaya_block.experts.local_experts.5.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.7.zaya_block.experts.local_experts.5.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.7.zaya_block.experts.local_experts.6.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.7.zaya_block.experts.local_experts.6.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.7.zaya_block.experts.local_experts.7.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.7.zaya_block.experts.local_experts.7.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.7.zaya_block.experts.local_experts.8.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.7.zaya_block.experts.local_experts.8.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.7.zaya_block.experts.local_experts.9.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.7.zaya_block.experts.local_experts.9.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.7.zaya_block.router.balancing_biases": "model-00001-of-00004.safetensors", + "model.layers.7.zaya_block.router.down_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.7.zaya_block.router.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.7.zaya_block.router.rmsnorm_eda.weight": "model-00001-of-00004.safetensors", + "model.layers.7.zaya_block.router.router_mlp.0.bias": "model-00001-of-00004.safetensors", + "model.layers.7.zaya_block.router.router_mlp.0.weight": "model-00001-of-00004.safetensors", + "model.layers.7.zaya_block.router.router_mlp.2.bias": "model-00001-of-00004.safetensors", + "model.layers.7.zaya_block.router.router_mlp.2.weight": "model-00001-of-00004.safetensors", + "model.layers.7.zaya_block.router.router_mlp.4.weight": "model-00001-of-00004.safetensors", + "model.layers.7.zaya_block.router.router_states_scale": "model-00001-of-00004.safetensors", + "model.layers.70.input_norm.weight": "model-00004-of-00004.safetensors", + "model.layers.70.res_scale.hidden_states_bias": "model-00004-of-00004.safetensors", + "model.layers.70.res_scale.hidden_states_scale": "model-00004-of-00004.safetensors", + "model.layers.70.res_scale.residual_bias": "model-00004-of-00004.safetensors", + "model.layers.70.res_scale.residual_scale": "model-00004-of-00004.safetensors", + "model.layers.70.self_attn.o_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.70.self_attn.qkv.conv_qk.0.bias": "model-00004-of-00004.safetensors", + "model.layers.70.self_attn.qkv.conv_qk.0.weight": "model-00004-of-00004.safetensors", + "model.layers.70.self_attn.qkv.conv_qk.1.bias": "model-00004-of-00004.safetensors", + "model.layers.70.self_attn.qkv.conv_qk.1.weight": "model-00004-of-00004.safetensors", + "model.layers.70.self_attn.qkv.linear_k.weight": "model-00004-of-00004.safetensors", + "model.layers.70.self_attn.qkv.linear_q.weight": "model-00004-of-00004.safetensors", + "model.layers.70.self_attn.qkv.temp": "model-00004-of-00004.safetensors", + "model.layers.70.self_attn.qkv.val_proj1.weight": "model-00004-of-00004.safetensors", + "model.layers.70.self_attn.qkv.val_proj2.weight": "model-00004-of-00004.safetensors", + "model.layers.71.input_norm.weight": "model-00004-of-00004.safetensors", + "model.layers.71.res_scale.hidden_states_bias": "model-00004-of-00004.safetensors", + "model.layers.71.res_scale.hidden_states_scale": "model-00004-of-00004.safetensors", + "model.layers.71.res_scale.residual_bias": "model-00004-of-00004.safetensors", + "model.layers.71.res_scale.residual_scale": "model-00004-of-00004.safetensors", + "model.layers.71.zaya_block.experts.local_experts.0.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.71.zaya_block.experts.local_experts.0.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.71.zaya_block.experts.local_experts.1.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.71.zaya_block.experts.local_experts.1.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.71.zaya_block.experts.local_experts.10.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.71.zaya_block.experts.local_experts.10.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.71.zaya_block.experts.local_experts.11.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.71.zaya_block.experts.local_experts.11.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.71.zaya_block.experts.local_experts.12.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.71.zaya_block.experts.local_experts.12.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.71.zaya_block.experts.local_experts.13.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.71.zaya_block.experts.local_experts.13.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.71.zaya_block.experts.local_experts.14.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.71.zaya_block.experts.local_experts.14.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.71.zaya_block.experts.local_experts.15.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.71.zaya_block.experts.local_experts.15.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.71.zaya_block.experts.local_experts.2.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.71.zaya_block.experts.local_experts.2.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.71.zaya_block.experts.local_experts.3.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.71.zaya_block.experts.local_experts.3.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.71.zaya_block.experts.local_experts.4.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.71.zaya_block.experts.local_experts.4.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.71.zaya_block.experts.local_experts.5.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.71.zaya_block.experts.local_experts.5.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.71.zaya_block.experts.local_experts.6.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.71.zaya_block.experts.local_experts.6.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.71.zaya_block.experts.local_experts.7.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.71.zaya_block.experts.local_experts.7.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.71.zaya_block.experts.local_experts.8.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.71.zaya_block.experts.local_experts.8.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.71.zaya_block.experts.local_experts.9.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.71.zaya_block.experts.local_experts.9.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.71.zaya_block.router.balancing_biases": "model-00004-of-00004.safetensors", + "model.layers.71.zaya_block.router.down_proj.bias": "model-00004-of-00004.safetensors", + "model.layers.71.zaya_block.router.down_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.71.zaya_block.router.rmsnorm_eda.weight": "model-00004-of-00004.safetensors", + "model.layers.71.zaya_block.router.router_mlp.0.bias": "model-00004-of-00004.safetensors", + "model.layers.71.zaya_block.router.router_mlp.0.weight": "model-00004-of-00004.safetensors", + "model.layers.71.zaya_block.router.router_mlp.2.bias": "model-00004-of-00004.safetensors", + "model.layers.71.zaya_block.router.router_mlp.2.weight": "model-00004-of-00004.safetensors", + "model.layers.71.zaya_block.router.router_mlp.4.weight": "model-00004-of-00004.safetensors", + "model.layers.71.zaya_block.router.router_states_scale": "model-00004-of-00004.safetensors", + "model.layers.72.input_norm.weight": "model-00004-of-00004.safetensors", + "model.layers.72.res_scale.hidden_states_bias": "model-00004-of-00004.safetensors", + "model.layers.72.res_scale.hidden_states_scale": "model-00004-of-00004.safetensors", + "model.layers.72.res_scale.residual_bias": "model-00004-of-00004.safetensors", + "model.layers.72.res_scale.residual_scale": "model-00004-of-00004.safetensors", + "model.layers.72.self_attn.o_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.72.self_attn.qkv.conv_qk.0.bias": "model-00004-of-00004.safetensors", + "model.layers.72.self_attn.qkv.conv_qk.0.weight": "model-00004-of-00004.safetensors", + "model.layers.72.self_attn.qkv.conv_qk.1.bias": "model-00004-of-00004.safetensors", + "model.layers.72.self_attn.qkv.conv_qk.1.weight": "model-00004-of-00004.safetensors", + "model.layers.72.self_attn.qkv.linear_k.weight": "model-00004-of-00004.safetensors", + "model.layers.72.self_attn.qkv.linear_q.weight": "model-00004-of-00004.safetensors", + "model.layers.72.self_attn.qkv.temp": "model-00004-of-00004.safetensors", + "model.layers.72.self_attn.qkv.val_proj1.weight": "model-00004-of-00004.safetensors", + "model.layers.72.self_attn.qkv.val_proj2.weight": "model-00004-of-00004.safetensors", + "model.layers.73.input_norm.weight": "model-00004-of-00004.safetensors", + "model.layers.73.res_scale.hidden_states_bias": "model-00004-of-00004.safetensors", + "model.layers.73.res_scale.hidden_states_scale": "model-00004-of-00004.safetensors", + "model.layers.73.res_scale.residual_bias": "model-00004-of-00004.safetensors", + "model.layers.73.res_scale.residual_scale": "model-00004-of-00004.safetensors", + "model.layers.73.zaya_block.experts.local_experts.0.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.73.zaya_block.experts.local_experts.0.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.73.zaya_block.experts.local_experts.1.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.73.zaya_block.experts.local_experts.1.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.73.zaya_block.experts.local_experts.10.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.73.zaya_block.experts.local_experts.10.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.73.zaya_block.experts.local_experts.11.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.73.zaya_block.experts.local_experts.11.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.73.zaya_block.experts.local_experts.12.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.73.zaya_block.experts.local_experts.12.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.73.zaya_block.experts.local_experts.13.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.73.zaya_block.experts.local_experts.13.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.73.zaya_block.experts.local_experts.14.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.73.zaya_block.experts.local_experts.14.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.73.zaya_block.experts.local_experts.15.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.73.zaya_block.experts.local_experts.15.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.73.zaya_block.experts.local_experts.2.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.73.zaya_block.experts.local_experts.2.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.73.zaya_block.experts.local_experts.3.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.73.zaya_block.experts.local_experts.3.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.73.zaya_block.experts.local_experts.4.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.73.zaya_block.experts.local_experts.4.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.73.zaya_block.experts.local_experts.5.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.73.zaya_block.experts.local_experts.5.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.73.zaya_block.experts.local_experts.6.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.73.zaya_block.experts.local_experts.6.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.73.zaya_block.experts.local_experts.7.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.73.zaya_block.experts.local_experts.7.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.73.zaya_block.experts.local_experts.8.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.73.zaya_block.experts.local_experts.8.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.73.zaya_block.experts.local_experts.9.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.73.zaya_block.experts.local_experts.9.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.73.zaya_block.router.balancing_biases": "model-00004-of-00004.safetensors", + "model.layers.73.zaya_block.router.down_proj.bias": "model-00004-of-00004.safetensors", + "model.layers.73.zaya_block.router.down_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.73.zaya_block.router.rmsnorm_eda.weight": "model-00004-of-00004.safetensors", + "model.layers.73.zaya_block.router.router_mlp.0.bias": "model-00004-of-00004.safetensors", + "model.layers.73.zaya_block.router.router_mlp.0.weight": "model-00004-of-00004.safetensors", + "model.layers.73.zaya_block.router.router_mlp.2.bias": "model-00004-of-00004.safetensors", + "model.layers.73.zaya_block.router.router_mlp.2.weight": "model-00004-of-00004.safetensors", + "model.layers.73.zaya_block.router.router_mlp.4.weight": "model-00004-of-00004.safetensors", + "model.layers.73.zaya_block.router.router_states_scale": "model-00004-of-00004.safetensors", + "model.layers.74.input_norm.weight": "model-00004-of-00004.safetensors", + "model.layers.74.res_scale.hidden_states_bias": "model-00004-of-00004.safetensors", + "model.layers.74.res_scale.hidden_states_scale": "model-00004-of-00004.safetensors", + "model.layers.74.res_scale.residual_bias": "model-00004-of-00004.safetensors", + "model.layers.74.res_scale.residual_scale": "model-00004-of-00004.safetensors", + "model.layers.74.self_attn.o_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.74.self_attn.qkv.conv_qk.0.bias": "model-00004-of-00004.safetensors", + "model.layers.74.self_attn.qkv.conv_qk.0.weight": "model-00004-of-00004.safetensors", + "model.layers.74.self_attn.qkv.conv_qk.1.bias": "model-00004-of-00004.safetensors", + "model.layers.74.self_attn.qkv.conv_qk.1.weight": "model-00004-of-00004.safetensors", + "model.layers.74.self_attn.qkv.linear_k.weight": "model-00004-of-00004.safetensors", + "model.layers.74.self_attn.qkv.linear_q.weight": "model-00004-of-00004.safetensors", + "model.layers.74.self_attn.qkv.temp": "model-00004-of-00004.safetensors", + "model.layers.74.self_attn.qkv.val_proj1.weight": "model-00004-of-00004.safetensors", + "model.layers.74.self_attn.qkv.val_proj2.weight": "model-00004-of-00004.safetensors", + "model.layers.75.input_norm.weight": "model-00004-of-00004.safetensors", + "model.layers.75.res_scale.hidden_states_bias": "model-00004-of-00004.safetensors", + "model.layers.75.res_scale.hidden_states_scale": "model-00004-of-00004.safetensors", + "model.layers.75.res_scale.residual_bias": "model-00004-of-00004.safetensors", + "model.layers.75.res_scale.residual_scale": "model-00004-of-00004.safetensors", + "model.layers.75.zaya_block.experts.local_experts.0.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.75.zaya_block.experts.local_experts.0.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.75.zaya_block.experts.local_experts.1.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.75.zaya_block.experts.local_experts.1.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.75.zaya_block.experts.local_experts.10.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.75.zaya_block.experts.local_experts.10.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.75.zaya_block.experts.local_experts.11.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.75.zaya_block.experts.local_experts.11.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.75.zaya_block.experts.local_experts.12.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.75.zaya_block.experts.local_experts.12.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.75.zaya_block.experts.local_experts.13.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.75.zaya_block.experts.local_experts.13.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.75.zaya_block.experts.local_experts.14.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.75.zaya_block.experts.local_experts.14.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.75.zaya_block.experts.local_experts.15.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.75.zaya_block.experts.local_experts.15.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.75.zaya_block.experts.local_experts.2.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.75.zaya_block.experts.local_experts.2.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.75.zaya_block.experts.local_experts.3.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.75.zaya_block.experts.local_experts.3.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.75.zaya_block.experts.local_experts.4.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.75.zaya_block.experts.local_experts.4.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.75.zaya_block.experts.local_experts.5.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.75.zaya_block.experts.local_experts.5.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.75.zaya_block.experts.local_experts.6.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.75.zaya_block.experts.local_experts.6.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.75.zaya_block.experts.local_experts.7.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.75.zaya_block.experts.local_experts.7.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.75.zaya_block.experts.local_experts.8.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.75.zaya_block.experts.local_experts.8.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.75.zaya_block.experts.local_experts.9.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.75.zaya_block.experts.local_experts.9.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.75.zaya_block.router.balancing_biases": "model-00004-of-00004.safetensors", + "model.layers.75.zaya_block.router.down_proj.bias": "model-00004-of-00004.safetensors", + "model.layers.75.zaya_block.router.down_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.75.zaya_block.router.rmsnorm_eda.weight": "model-00004-of-00004.safetensors", + "model.layers.75.zaya_block.router.router_mlp.0.bias": "model-00004-of-00004.safetensors", + "model.layers.75.zaya_block.router.router_mlp.0.weight": "model-00004-of-00004.safetensors", + "model.layers.75.zaya_block.router.router_mlp.2.bias": "model-00004-of-00004.safetensors", + "model.layers.75.zaya_block.router.router_mlp.2.weight": "model-00004-of-00004.safetensors", + "model.layers.75.zaya_block.router.router_mlp.4.weight": "model-00004-of-00004.safetensors", + "model.layers.75.zaya_block.router.router_states_scale": "model-00004-of-00004.safetensors", + "model.layers.76.input_norm.weight": "model-00004-of-00004.safetensors", + "model.layers.76.res_scale.hidden_states_bias": "model-00004-of-00004.safetensors", + "model.layers.76.res_scale.hidden_states_scale": "model-00004-of-00004.safetensors", + "model.layers.76.res_scale.residual_bias": "model-00004-of-00004.safetensors", + "model.layers.76.res_scale.residual_scale": "model-00004-of-00004.safetensors", + "model.layers.76.self_attn.o_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.76.self_attn.qkv.conv_qk.0.bias": "model-00004-of-00004.safetensors", + "model.layers.76.self_attn.qkv.conv_qk.0.weight": "model-00004-of-00004.safetensors", + "model.layers.76.self_attn.qkv.conv_qk.1.bias": "model-00004-of-00004.safetensors", + "model.layers.76.self_attn.qkv.conv_qk.1.weight": "model-00004-of-00004.safetensors", + "model.layers.76.self_attn.qkv.linear_k.weight": "model-00004-of-00004.safetensors", + "model.layers.76.self_attn.qkv.linear_q.weight": "model-00004-of-00004.safetensors", + "model.layers.76.self_attn.qkv.temp": "model-00004-of-00004.safetensors", + "model.layers.76.self_attn.qkv.val_proj1.weight": "model-00004-of-00004.safetensors", + "model.layers.76.self_attn.qkv.val_proj2.weight": "model-00004-of-00004.safetensors", + "model.layers.77.input_norm.weight": "model-00004-of-00004.safetensors", + "model.layers.77.res_scale.hidden_states_bias": "model-00004-of-00004.safetensors", + "model.layers.77.res_scale.hidden_states_scale": "model-00004-of-00004.safetensors", + "model.layers.77.res_scale.residual_bias": "model-00004-of-00004.safetensors", + "model.layers.77.res_scale.residual_scale": "model-00004-of-00004.safetensors", + "model.layers.77.zaya_block.experts.local_experts.0.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.77.zaya_block.experts.local_experts.0.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.77.zaya_block.experts.local_experts.1.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.77.zaya_block.experts.local_experts.1.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.77.zaya_block.experts.local_experts.10.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.77.zaya_block.experts.local_experts.10.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.77.zaya_block.experts.local_experts.11.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.77.zaya_block.experts.local_experts.11.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.77.zaya_block.experts.local_experts.12.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.77.zaya_block.experts.local_experts.12.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.77.zaya_block.experts.local_experts.13.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.77.zaya_block.experts.local_experts.13.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.77.zaya_block.experts.local_experts.14.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.77.zaya_block.experts.local_experts.14.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.77.zaya_block.experts.local_experts.15.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.77.zaya_block.experts.local_experts.15.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.77.zaya_block.experts.local_experts.2.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.77.zaya_block.experts.local_experts.2.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.77.zaya_block.experts.local_experts.3.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.77.zaya_block.experts.local_experts.3.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.77.zaya_block.experts.local_experts.4.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.77.zaya_block.experts.local_experts.4.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.77.zaya_block.experts.local_experts.5.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.77.zaya_block.experts.local_experts.5.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.77.zaya_block.experts.local_experts.6.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.77.zaya_block.experts.local_experts.6.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.77.zaya_block.experts.local_experts.7.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.77.zaya_block.experts.local_experts.7.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.77.zaya_block.experts.local_experts.8.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.77.zaya_block.experts.local_experts.8.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.77.zaya_block.experts.local_experts.9.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.77.zaya_block.experts.local_experts.9.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.77.zaya_block.router.balancing_biases": "model-00004-of-00004.safetensors", + "model.layers.77.zaya_block.router.down_proj.bias": "model-00004-of-00004.safetensors", + "model.layers.77.zaya_block.router.down_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.77.zaya_block.router.rmsnorm_eda.weight": "model-00004-of-00004.safetensors", + "model.layers.77.zaya_block.router.router_mlp.0.bias": "model-00004-of-00004.safetensors", + "model.layers.77.zaya_block.router.router_mlp.0.weight": "model-00004-of-00004.safetensors", + "model.layers.77.zaya_block.router.router_mlp.2.bias": "model-00004-of-00004.safetensors", + "model.layers.77.zaya_block.router.router_mlp.2.weight": "model-00004-of-00004.safetensors", + "model.layers.77.zaya_block.router.router_mlp.4.weight": "model-00004-of-00004.safetensors", + "model.layers.77.zaya_block.router.router_states_scale": "model-00004-of-00004.safetensors", + "model.layers.78.input_norm.weight": "model-00004-of-00004.safetensors", + "model.layers.78.res_scale.hidden_states_bias": "model-00004-of-00004.safetensors", + "model.layers.78.res_scale.hidden_states_scale": "model-00004-of-00004.safetensors", + "model.layers.78.res_scale.residual_bias": "model-00004-of-00004.safetensors", + "model.layers.78.res_scale.residual_scale": "model-00004-of-00004.safetensors", + "model.layers.78.self_attn.o_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.78.self_attn.qkv.conv_qk.0.bias": "model-00004-of-00004.safetensors", + "model.layers.78.self_attn.qkv.conv_qk.0.weight": "model-00004-of-00004.safetensors", + "model.layers.78.self_attn.qkv.conv_qk.1.bias": "model-00004-of-00004.safetensors", + "model.layers.78.self_attn.qkv.conv_qk.1.weight": "model-00004-of-00004.safetensors", + "model.layers.78.self_attn.qkv.linear_k.weight": "model-00004-of-00004.safetensors", + "model.layers.78.self_attn.qkv.linear_q.weight": "model-00004-of-00004.safetensors", + "model.layers.78.self_attn.qkv.temp": "model-00004-of-00004.safetensors", + "model.layers.78.self_attn.qkv.val_proj1.weight": "model-00004-of-00004.safetensors", + "model.layers.78.self_attn.qkv.val_proj2.weight": "model-00004-of-00004.safetensors", + "model.layers.79.input_norm.weight": "model-00004-of-00004.safetensors", + "model.layers.79.res_scale.hidden_states_bias": "model-00004-of-00004.safetensors", + "model.layers.79.res_scale.hidden_states_scale": "model-00004-of-00004.safetensors", + "model.layers.79.res_scale.residual_bias": "model-00004-of-00004.safetensors", + "model.layers.79.res_scale.residual_scale": "model-00004-of-00004.safetensors", + "model.layers.79.zaya_block.experts.local_experts.0.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.79.zaya_block.experts.local_experts.0.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.79.zaya_block.experts.local_experts.1.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.79.zaya_block.experts.local_experts.1.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.79.zaya_block.experts.local_experts.10.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.79.zaya_block.experts.local_experts.10.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.79.zaya_block.experts.local_experts.11.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.79.zaya_block.experts.local_experts.11.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.79.zaya_block.experts.local_experts.12.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.79.zaya_block.experts.local_experts.12.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.79.zaya_block.experts.local_experts.13.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.79.zaya_block.experts.local_experts.13.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.79.zaya_block.experts.local_experts.14.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.79.zaya_block.experts.local_experts.14.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.79.zaya_block.experts.local_experts.15.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.79.zaya_block.experts.local_experts.15.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.79.zaya_block.experts.local_experts.2.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.79.zaya_block.experts.local_experts.2.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.79.zaya_block.experts.local_experts.3.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.79.zaya_block.experts.local_experts.3.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.79.zaya_block.experts.local_experts.4.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.79.zaya_block.experts.local_experts.4.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.79.zaya_block.experts.local_experts.5.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.79.zaya_block.experts.local_experts.5.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.79.zaya_block.experts.local_experts.6.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.79.zaya_block.experts.local_experts.6.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.79.zaya_block.experts.local_experts.7.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.79.zaya_block.experts.local_experts.7.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.79.zaya_block.experts.local_experts.8.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.79.zaya_block.experts.local_experts.8.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.79.zaya_block.experts.local_experts.9.linear_fc1.weight": "model-00004-of-00004.safetensors", + "model.layers.79.zaya_block.experts.local_experts.9.linear_fc2.weight": "model-00004-of-00004.safetensors", + "model.layers.79.zaya_block.router.balancing_biases": "model-00004-of-00004.safetensors", + "model.layers.79.zaya_block.router.down_proj.bias": "model-00004-of-00004.safetensors", + "model.layers.79.zaya_block.router.down_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.79.zaya_block.router.rmsnorm_eda.weight": "model-00004-of-00004.safetensors", + "model.layers.79.zaya_block.router.router_mlp.0.bias": "model-00004-of-00004.safetensors", + "model.layers.79.zaya_block.router.router_mlp.0.weight": "model-00004-of-00004.safetensors", + "model.layers.79.zaya_block.router.router_mlp.2.bias": "model-00004-of-00004.safetensors", + "model.layers.79.zaya_block.router.router_mlp.2.weight": "model-00004-of-00004.safetensors", + "model.layers.79.zaya_block.router.router_mlp.4.weight": "model-00004-of-00004.safetensors", + "model.layers.79.zaya_block.router.router_states_scale": "model-00004-of-00004.safetensors", + "model.layers.8.input_norm.weight": "model-00001-of-00004.safetensors", + "model.layers.8.res_scale.hidden_states_bias": "model-00001-of-00004.safetensors", + "model.layers.8.res_scale.hidden_states_scale": "model-00001-of-00004.safetensors", + "model.layers.8.res_scale.residual_bias": "model-00001-of-00004.safetensors", + "model.layers.8.res_scale.residual_scale": "model-00001-of-00004.safetensors", + "model.layers.8.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.8.self_attn.qkv.conv_qk.0.bias": "model-00001-of-00004.safetensors", + "model.layers.8.self_attn.qkv.conv_qk.0.weight": "model-00001-of-00004.safetensors", + "model.layers.8.self_attn.qkv.conv_qk.1.bias": "model-00001-of-00004.safetensors", + "model.layers.8.self_attn.qkv.conv_qk.1.weight": "model-00001-of-00004.safetensors", + "model.layers.8.self_attn.qkv.linear_k.weight": "model-00001-of-00004.safetensors", + "model.layers.8.self_attn.qkv.linear_q.weight": "model-00001-of-00004.safetensors", + "model.layers.8.self_attn.qkv.temp": "model-00001-of-00004.safetensors", + "model.layers.8.self_attn.qkv.val_proj1.weight": "model-00001-of-00004.safetensors", + "model.layers.8.self_attn.qkv.val_proj2.weight": "model-00001-of-00004.safetensors", + "model.layers.9.input_norm.weight": "model-00001-of-00004.safetensors", + "model.layers.9.res_scale.hidden_states_bias": "model-00001-of-00004.safetensors", + "model.layers.9.res_scale.hidden_states_scale": "model-00001-of-00004.safetensors", + "model.layers.9.res_scale.residual_bias": "model-00001-of-00004.safetensors", + "model.layers.9.res_scale.residual_scale": "model-00001-of-00004.safetensors", + "model.layers.9.zaya_block.experts.local_experts.0.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.9.zaya_block.experts.local_experts.0.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.9.zaya_block.experts.local_experts.1.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.9.zaya_block.experts.local_experts.1.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.9.zaya_block.experts.local_experts.10.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.9.zaya_block.experts.local_experts.10.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.9.zaya_block.experts.local_experts.11.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.9.zaya_block.experts.local_experts.11.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.9.zaya_block.experts.local_experts.12.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.9.zaya_block.experts.local_experts.12.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.9.zaya_block.experts.local_experts.13.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.9.zaya_block.experts.local_experts.13.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.9.zaya_block.experts.local_experts.14.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.9.zaya_block.experts.local_experts.14.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.9.zaya_block.experts.local_experts.15.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.9.zaya_block.experts.local_experts.15.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.9.zaya_block.experts.local_experts.2.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.9.zaya_block.experts.local_experts.2.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.9.zaya_block.experts.local_experts.3.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.9.zaya_block.experts.local_experts.3.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.9.zaya_block.experts.local_experts.4.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.9.zaya_block.experts.local_experts.4.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.9.zaya_block.experts.local_experts.5.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.9.zaya_block.experts.local_experts.5.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.9.zaya_block.experts.local_experts.6.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.9.zaya_block.experts.local_experts.6.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.9.zaya_block.experts.local_experts.7.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.9.zaya_block.experts.local_experts.7.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.9.zaya_block.experts.local_experts.8.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.9.zaya_block.experts.local_experts.8.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.9.zaya_block.experts.local_experts.9.linear_fc1.weight": "model-00001-of-00004.safetensors", + "model.layers.9.zaya_block.experts.local_experts.9.linear_fc2.weight": "model-00001-of-00004.safetensors", + "model.layers.9.zaya_block.router.balancing_biases": "model-00001-of-00004.safetensors", + "model.layers.9.zaya_block.router.down_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.9.zaya_block.router.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.9.zaya_block.router.rmsnorm_eda.weight": "model-00001-of-00004.safetensors", + "model.layers.9.zaya_block.router.router_mlp.0.bias": "model-00001-of-00004.safetensors", + "model.layers.9.zaya_block.router.router_mlp.0.weight": "model-00001-of-00004.safetensors", + "model.layers.9.zaya_block.router.router_mlp.2.bias": "model-00001-of-00004.safetensors", + "model.layers.9.zaya_block.router.router_mlp.2.weight": "model-00001-of-00004.safetensors", + "model.layers.9.zaya_block.router.router_mlp.4.weight": "model-00001-of-00004.safetensors", + "model.layers.9.zaya_block.router.router_states_scale": "model-00001-of-00004.safetensors", + "model.res_scale.hidden_states_bias": "model-00004-of-00004.safetensors", + "model.res_scale.hidden_states_scale": "model-00004-of-00004.safetensors", + "model.res_scale.residual_bias": "model-00004-of-00004.safetensors", + "model.res_scale.residual_scale": "model-00004-of-00004.safetensors" + } +} \ No newline at end of file