checkpoint-190000/config.json

{
  "ar_steps": 1,
  "architectures": [
    "DiffVLMBaseline"
  ],
  "attention_dropout": 0.0,
  "bos_token_id": 151643,
  "condition_layer": -1,
  "eos_token_id": 151645,
  "hidden_act": "silu",
  "hidden_size": 1536,
  "image_token_id": 151655,
  "img_cross_attention_dim": 2048,
  "img_diffuser_depth": 2,
  "img_ffn_dim_multiplier": null,
  "img_hidden_size": 1536,
  "img_multiple_of": 256,
  "img_norm_eps": 1e-05,
  "img_num_attention_heads": 12,
  "img_num_kv_heads": 12,
  "img_qk_norm": true,
  "in_channels": 32,
  "initializer_range": 0.02,
  "inject_img_diffuser": false,
  "input_size": 32,
  "intermediate_size": 8960,
  "layer_group_size": 7,
  "layerwise_start_idx": 0,
  "lora_alpha": 128,
  "lora_bias": "none",
  "lora_dropout": 0.05,
  "lora_enable": false,
  "lora_r": 64,
  "max_position_embeddings": 32768,
  "max_window_layers": 28,
  "model_type": "qwen2_vl",
  "non_linearity": 1,
  "norm_elementwise_affine": true,
  "num_attention_heads": 12,
  "num_hidden_layers": 28,
  "num_key_value_heads": 2,
  "patch_size": 1,
  "repa_coeff": 0.1,
  "repa_layers": "2",
  "repa_shared": false,
  "rms_norm_eps": 1e-06,
  "rope_scaling": {
    "mrope_section": [
      16,
      24,
      24
    ],
    "rope_type": "default",
    "type": "default"
  },
  "rope_theta": 1000000.0,
  "sample_size": 128,
  "sampling_steps": 28,
  "sliding_window": null,
  "tie_word_embeddings": true,
  "torch_dtype": "bfloat16",
  "transformers_version": "4.47.0",
  "use_cache": true,
  "use_repa": false,
  "use_residual_attn": false,
  "use_sliding_window": false,
  "vae_path": "mit-han-lab/dc-ae-f32c32-in-1.0-diffusers",
  "video_token_id": 151656,
  "vision_config": {
    "hidden_size": 1536,
    "in_chans": 3,
    "model_type": "qwen2_vl",
    "spatial_patch_size": 14
  },
  "vision_end_token_id": 151653,
  "vision_start_token_id": 151652,
  "vision_token_id": 151654,
  "vocab_size": 151936
}