| { |
| "architectures": [ |
| "Ernie4_5_VLMoeModel" |
| ], |
| "dtype": "float32", |
| "image_end_token_id": 6, |
| "image_start_token_id": 5, |
| "image_token_id": 7, |
| "model_type": "ernie4_5_vl_moe", |
| "text_config": { |
| "bos_token_id": null, |
| "eos_token_id": null, |
| "hidden_act": "silu", |
| "hidden_size": 16, |
| "initializer_range": 0.02, |
| "intermediate_size": 32, |
| "max_position_embeddings": 512, |
| "mlp_layer_types": [ |
| "dense", |
| "sparse" |
| ], |
| "model_type": "ernie4_5_vl_moe_text", |
| "moe_intermediate_size": [ |
| 32, |
| 32 |
| ], |
| "moe_k": 2, |
| "moe_norm_min": 1e-12, |
| "moe_num_experts": 8, |
| "moe_num_shared_experts": 2, |
| "num_attention_heads": 2, |
| "num_hidden_layers": 2, |
| "num_key_value_heads": 1, |
| "output_router_logits": false, |
| "pad_token_id": null, |
| "rms_norm_eps": 1e-05, |
| "rope_parameters": { |
| "mrope_section": [ |
| 1, |
| 1, |
| 2 |
| ], |
| "rope_theta": 500000.0, |
| "rope_type": "default", |
| "type": "default" |
| }, |
| "router_aux_loss_coef": 0.001, |
| "tie_word_embeddings": true, |
| "use_bias": false, |
| "use_cache": true, |
| "vocab_size": 103368 |
| }, |
| "tie_word_embeddings": true, |
| "transformers_version": "5.16.0.dev0", |
| "video_end_token_id": 4, |
| "video_start_token_id": 3, |
| "video_token_id": 8, |
| "vision_config": { |
| "depth": 2, |
| "hidden_act": "silu", |
| "hidden_size": 32, |
| "in_channels": 3, |
| "initializer_range": 0.02, |
| "intermediate_size": 32, |
| "model_type": "ernie4_5_vl_moe_vision", |
| "num_heads": 2, |
| "patch_size": 14, |
| "rms_norm_eps": 1e-06, |
| "spatial_merge_size": 1, |
| "temporal_merge_size": 2 |
| } |
| } |
|
|
