WORLD_MODEL_POSTTRAIN/config.json

{
  "_name_or_path": "/share/project/yuqi.wang/OmniSim/pretrain/Emu3-Base",
  "action_config": {
    "_attn_implementation": "flash_attention_2",
    "action_dim": 7,
    "attention_dropout": 0.0,
    "hidden_act": "silu",
    "hidden_size": 4096,
    "initializer_range": 0.02,
    "intermediate_size": 2048,
    "max_position_embeddings": 5120,
    "num_attention_heads": 32,
    "num_hidden_layers": 2,
    "num_key_value_heads": 8,
    "pad_token_id": 151643,
    "pretraining_tp": 1,
    "rms_norm_eps": 1e-05,
    "rope_scaling": null,
    "rope_theta": 10000.0,
    "tie_word_embeddings": false,
    "use_cache": true,
    "vision_loss_weight": 5.0
  },
  "action_dim": 7,
  "action_experts": false,
  "architectures": [
    "Emu3MoE"
  ],
  "attention_dropout": 0.1,
  "boa_token_id": 151844,
  "boi_token_id": 151852,
  "bos_token_id": 151849,
  "bov_token_id": 151854,
  "eof_token_id": 151847,
  "eoi_token_id": 151853,
  "eol_token_id": 151846,
  "eos_token_id": 151850,
  "eov_token_id": 184621,
  "hidden_act": "silu",
  "hidden_size": 4096,
  "image_area": 262144,
  "img_token_id": 151851,
  "initializer_range": 0.02,
  "intermediate_size": 14336,
  "max_position_embeddings": 6400,
  "model_type": "Emu3",
  "num_attention_heads": 32,
  "num_hidden_layers": 32,
  "num_key_value_heads": 8,
  "pad_token_id": 151643,
  "pretraining_tp": 1,
  "qkv_bias": false,
  "rms_norm_eps": 1e-05,
  "rope_scaling": null,
  "rope_theta": 1000000.0,
  "tie_word_embeddings": false,
  "torch_dtype": "bfloat16",
  "transformers_version": "4.44.0",
  "use_cache": true,
  "vision_loss_weight": 1.0,
  "vocab_size": 184622
}