| { |
| "architectures": [ |
| "PerceptionLMModel" |
| ], |
| "dtype": "float32", |
| "image_token_id": 0, |
| "model_type": "perception_lm", |
| "projector_pooling_ratio": 1, |
| "text_config": { |
| "attention_bias": false, |
| "attention_dropout": 0.0, |
| "attention_probs_dropout_prob": 0.1, |
| "bos_token_id": 128000, |
| "eos_token_id": 128009, |
| "head_dim": 8, |
| "hidden_act": "gelu", |
| "hidden_dropout_prob": 0.1, |
| "hidden_size": 32, |
| "id2label": { |
| "0": "LABEL_0", |
| "1": "LABEL_1", |
| "2": "LABEL_2" |
| }, |
| "initializer_range": 0.02, |
| "intermediate_size": 37, |
| "is_training": true, |
| "label2id": { |
| "LABEL_0": 0, |
| "LABEL_1": 1, |
| "LABEL_2": 2 |
| }, |
| "max_position_embeddings": 512, |
| "mlp_bias": false, |
| "model_type": "llama", |
| "num_attention_heads": 4, |
| "num_choices": 4, |
| "num_hidden_layers": 2, |
| "num_key_value_heads": 4, |
| "pad_token_id": 128001, |
| "pretraining_tp": 1, |
| "rms_norm_eps": 1e-06, |
| "rope_parameters": { |
| "rope_theta": 10000.0, |
| "rope_type": "default" |
| }, |
| "seq_length": 7, |
| "tie_word_embeddings": false, |
| "type_sequence_label_size": 2, |
| "type_vocab_size": 16, |
| "use_cache": true, |
| "use_input_mask": true, |
| "use_labels": true, |
| "use_token_type_ids": false, |
| "vocab_size": 128256 |
| }, |
| "tie_word_embeddings": true, |
| "transformers_version": "5.16.0.dev0", |
| "video_token_id": 2, |
| "vision_config": { |
| "architecture": "vit_pe_core_large_patch14_336", |
| "do_pooling": true, |
| "initializer_range": 0.02, |
| "label_names": [ |
| "LABEL_0", |
| "LABEL_1" |
| ], |
| "model_args": { |
| "depth": 2, |
| "embed_dim": 64, |
| "global_pool": "", |
| "img_size": [ |
| 14, |
| 14 |
| ], |
| "init_values": 0.1, |
| "ref_feat_shape": [ |
| 1, |
| 1 |
| ], |
| "use_post_transformer_norm": false |
| }, |
| "model_type": "timm_wrapper", |
| "num_classes": 2 |
| }, |
| "vision_use_cls_token": true |
| } |
|
|
