| { | |
| "_name_or_path": "dandelin/vilt-b32-mlm", | |
| "architectures": [ | |
| "ViltForQuestionAnswering" | |
| ], | |
| "attention_probs_dropout_prob": 0.0, | |
| "hidden_act": "gelu", | |
| "hidden_dropout_prob": 0.0, | |
| "hidden_size": 768, | |
| "id2label": { | |
| "0": "10", | |
| "1": "walking", | |
| "2": "watching", | |
| "3": "1", | |
| "4": "yes", | |
| "5": "at table", | |
| "6": "gray", | |
| "7": "picnic table", | |
| "8": "green", | |
| "9": "4", | |
| "10": "shrimp", | |
| "11": "brown", | |
| "12": "tired", | |
| "13": "chopsticks", | |
| "14": "resting", | |
| "15": "down", | |
| "16": "crossing", | |
| "17": "table", | |
| "18": "ice cream", | |
| "19": "boy", | |
| "20": "skateboard", | |
| "21": "air", | |
| "22": "no" | |
| }, | |
| "image_size": 384, | |
| "initializer_range": 0.02, | |
| "intermediate_size": 3072, | |
| "label2id": { | |
| "1": 3, | |
| "10": 0, | |
| "4": 9, | |
| "air": 21, | |
| "at table": 5, | |
| "boy": 19, | |
| "brown": 11, | |
| "chopsticks": 13, | |
| "crossing": 16, | |
| "down": 15, | |
| "gray": 6, | |
| "green": 8, | |
| "ice cream": 18, | |
| "no": 22, | |
| "picnic table": 7, | |
| "resting": 14, | |
| "shrimp": 10, | |
| "skateboard": 20, | |
| "table": 17, | |
| "tired": 12, | |
| "walking": 1, | |
| "watching": 2, | |
| "yes": 4 | |
| }, | |
| "layer_norm_eps": 1e-12, | |
| "max_image_length": -1, | |
| "max_position_embeddings": 40, | |
| "modality_type_vocab_size": 2, | |
| "model_type": "vilt", | |
| "num_attention_heads": 12, | |
| "num_channels": 3, | |
| "num_hidden_layers": 12, | |
| "num_images": -1, | |
| "patch_size": 32, | |
| "qkv_bias": true, | |
| "tie_word_embeddings": false, | |
| "torch_dtype": "float32", | |
| "transformers_version": "4.35.2", | |
| "type_vocab_size": 2, | |
| "vocab_size": 30522 | |
| } | |