transformer/config.json

{
  "_class_name": "QwenImageTransformer2DModel",
  "_diffusers_version": "0.35.0.dev0",
  "_name_or_path": "/fsx/sayak/.cache/models--Qwen--Qwen-Image/snapshots/b0d9cbe58a813c316e7107e15310a4347b223020/transformer",
  "attention_head_dim": 128,
  "axes_dims_rope": [
    16,
    56,
    56
  ],
  "guidance_embeds": false,
  "in_channels": 64,
  "joint_attention_dim": 3584,
  "num_attention_heads": 24,
  "num_layers": 60,
  "out_channels": 16,
  "patch_size": 2,
  "pooled_projection_dim": 768,
  "quantization_config": {
    "_load_in_4bit": true,
    "_load_in_8bit": false,
    "bnb_4bit_compute_dtype": "bfloat16",
    "bnb_4bit_quant_storage": "uint8",
    "bnb_4bit_quant_type": "nf4",
    "bnb_4bit_use_double_quant": false,
    "llm_int8_enable_fp32_cpu_offload": false,
    "llm_int8_has_fp16_weight": false,
    "llm_int8_skip_modules": [
      "time_text_embedimg_in",
      "norm_out",
      "proj_out",
      "img_mod",
      "txt_mod"
    ],
    "llm_int8_threshold": 6.0,
    "load_in_4bit": true,
    "load_in_8bit": false,
    "quant_method": "bitsandbytes"
  }
}