| { |
| "architectures": [ |
| "MiniCPMV4_6Model" |
| ], |
| "downsample_mode": "16x", |
| "drop_vision_last_layer": false, |
| "dtype": "float32", |
| "image_size": 32, |
| "image_token_id": 100, |
| "insert_layer_id": 0, |
| "merge_kernel_size": [ |
| 2, |
| 2 |
| ], |
| "merger_times": 1, |
| "model_type": "minicpmv4_6", |
| "patch_size": 8, |
| "text_config": { |
| "attention_bias": false, |
| "attention_dropout": 0.0, |
| "bos_token_id": 248045, |
| "eos_token_id": 248046, |
| "head_dim": 16, |
| "hidden_act": "silu", |
| "hidden_size": 64, |
| "initializer_range": 0.02, |
| "intermediate_size": 37, |
| "layer_types": [ |
| "full_attention", |
| "linear_attention" |
| ], |
| "linear_conv_kernel_dim": 2, |
| "linear_key_head_dim": 16, |
| "linear_num_key_heads": 4, |
| "linear_num_value_heads": 8, |
| "linear_value_head_dim": 16, |
| "max_position_embeddings": 512, |
| "model_type": "qwen3_5_text", |
| "num_attention_heads": 4, |
| "num_hidden_layers": 2, |
| "num_key_value_heads": 2, |
| "pad_token_id": 248044, |
| "partial_rotary_factor": 0.25, |
| "rms_norm_eps": 1e-06, |
| "rope_parameters": { |
| "partial_rotary_factor": 0.25, |
| "rope_theta": 10000, |
| "rope_type": "default" |
| }, |
| "tie_word_embeddings": true, |
| "use_cache": true, |
| "vocab_size": 248094 |
| }, |
| "tie_word_embeddings": false, |
| "transformers_version": "5.16.0.dev0", |
| "video_token_id": null, |
| "vision_config": { |
| "attention_dropout": 0.0, |
| "hidden_act": "gelu_pytorch_tanh", |
| "hidden_size": 64, |
| "image_size": 32, |
| "insert_layer_id": 0, |
| "intermediate_size": 128, |
| "layer_norm_eps": 1e-06, |
| "model_type": "minicpmv4_6_vision", |
| "num_attention_heads": 4, |
| "num_channels": 3, |
| "num_hidden_layers": 2, |
| "patch_size": 8, |
| "window_kernel_size": [ |
| 2, |
| 2 |
| ] |
| } |
| } |
|
|