| { |
| "add_classification_head": false, |
| "architectures": [ |
| "UniComForCausalMM" |
| ], |
| "attention_bias": true, |
| "attention_dropout": 0.0, |
| "attention_head_dim": 128, |
| "attention_o_proj_bias": false, |
| "auto_map": { |
| "AutoConfig": "configuration_unicom.UniComConfig", |
| "AutoModel": "modeling_unicom.UniComModel", |
| "AutoModelForCausalLM": "modeling_unicom.UniComForCausalMM" |
| }, |
| "bos_token_id": 1, |
| "cfg_distilled": false, |
| "cla_share_factor": 1, |
| "class_num": 0, |
| "compression_checkpoint_path": null, |
| "cond_image_type": "vit", |
| "cond_token_attn_type": "full", |
| "dense_list": [ |
| 3584, |
| 0 |
| ], |
| "eod_token_id": 3, |
| "eos_token_id": 2, |
| "hidden_act": "silu", |
| "hidden_size": 3584, |
| "im_end_id": 5, |
| "im_newline_id": 11, |
| "im_start_id": 4, |
| "image_base_size": 1024, |
| "image_token_id": 8, |
| "initializer_range": 0.02, |
| "intermediate_size": 18944, |
| "kv_lora_rank": 512, |
| "mask_init_id": 12, |
| "max_image_tokens": 1024, |
| "max_position_embeddings": 32768, |
| "mlp_bias": false, |
| "model_type": "UniCom", |
| "model_version": "UniCom-Instruct", |
| "norm_type": "hf_rms", |
| "num_attention_heads": 28, |
| "num_hidden_layers": 28, |
| "num_key_value_heads": 4, |
| "pad_id": -1, |
| "pad_token_id": 0, |
| "pool_type": "last", |
| "pretraining_tp": 1, |
| "q_lora_rank": 1536, |
| "qk_nope_head_dim": 128, |
| "qk_rope_head_dim": 64, |
| "rms_norm_eps": 1e-06, |
| "rope_scaling": null, |
| "rope_theta": 1000000.0, |
| "rope_type": "default", |
| "siglip_downsample_factor": [ |
| 16, |
| 16 |
| ], |
| "siglip_latent_dim": 64, |
| "siglip_original_dim": 1152, |
| "text_end_id": 7, |
| "text_start_id": 6, |
| "tie_word_embeddings": false, |
| "transformers_version": "4.57.1", |
| "use_cache": true, |
| "use_cla": false, |
| "use_compile": false, |
| "use_meanflow": false, |
| "use_mla": false, |
| "use_qk_norm": false, |
| "use_rotary_pos_emb": true, |
| "v_head_dim": 128, |
| "video_end_id": 10, |
| "video_start_id": 9, |
| "vit": { |
| "_attn_implementation": "sdpa", |
| "attention_dropout": 0.0, |
| "hidden_act": "gelu_pytorch_tanh", |
| "hidden_size": 1152, |
| "intermediate_size": 4304, |
| "layer_norm_eps": 1e-06, |
| "num_attention_heads": 16, |
| "num_channels": 3, |
| "num_hidden_layers": 27, |
| "num_patches": 256, |
| "output_attentions": false, |
| "output_hidden_states": false, |
| "patch_size": 16, |
| "torch_dtype": "float32", |
| "use_return_dict": true |
| }, |
| "vit_aligner": { |
| "depth": 2, |
| "input_dim": 1216, |
| "n_embed": 3584, |
| "projector_type": "mlp_gelu", |
| "torch_dtype": "float32" |
| }, |
| "vit_processor": { |
| "do_normalize": true, |
| "do_rescale": true, |
| "do_resize": true, |
| "image_mean": [ |
| 0.5, |
| 0.5, |
| 0.5 |
| ], |
| "image_processor_type": "Siglip2ImageProcessorFast", |
| "image_std": [ |
| 0.5, |
| 0.5, |
| 0.5 |
| ], |
| "max_num_patches": 1024, |
| "patch_size": 16, |
| "resample": 2, |
| "rescale_factor": 0.00392156862745098 |
| }, |
| "vit_type": "siglip2-so400m-patch16-naflex", |
| "vocab_size": 153768 |
| } |
|
|
