config.json

{
    "architectures": [
        "JambaForCausalLM"
    ],
    "attention_dropout": 0.0,
    "attn_layer_offset": 7,
    "attn_layer_period": 14,
    "bos_token_id": 1,
    "dtype": "bfloat16",
    "eos_token_id": [
        2,
        519
    ],
    "expert_layer_offset": 1,
    "expert_layer_period": 2,
    "hidden_act": "silu",
    "hidden_size": 2560,
    "initializer_range": 0.02,
    "intermediate_size": 8192,
    "mamba_conv_bias": true,
    "mamba_d_conv": 4,
    "mamba_d_state": 16,
    "mamba_dt_rank": 160,
    "mamba_expand": 2,
    "mamba_proj_bias": false,
    "max_position_embeddings": 262144,
    "model_type": "jamba",
    "num_attention_heads": 20,
    "num_experts": 1,
    "num_experts_per_tok": 1,
    "num_hidden_layers": 28,
    "num_key_value_heads": 1,
    "num_logits_to_keep": 1,
    "output_router_logits": false,
    "pad_token_id": 0,
    "quantization": {
        "group_size": 64,
        "bits": 4,
        "mode": "affine"
    },
    "quantization_config": {
        "group_size": 64,
        "bits": 4,
        "mode": "affine"
    },
    "rms_norm_eps": 1e-06,
    "router_aux_loss_coef": 0.001,
    "sliding_window": null,
    "tie_word_embeddings": true,
    "transformers_version": "4.56.1",
    "use_cache": true,
    "use_mamba_kernels": true,
    "vocab_size": 65536
}