| { | |
| "_name_or_path": "/scratch/1/user/classanc/mlm-flops/ECIR/V14_MSMARCO_2173288", | |
| "adapters": { | |
| "adapters": {}, | |
| "config_map": {}, | |
| "fusion_config_map": {}, | |
| "fusions": {} | |
| }, | |
| "architectures": [ | |
| "BertModel" | |
| ], | |
| "attention_probs_dropout_prob": 0.1, | |
| "bos_token_id": 0, | |
| "classifier_dropout": null, | |
| "eos_token_id": 2, | |
| "gradient_checkpointing": false, | |
| "hidden_act": "gelu", | |
| "hidden_dropout_prob": 0.1, | |
| "hidden_size": 768, | |
| "initializer_range": 0.02, | |
| "intermediate_size": 3072, | |
| "layer_norm_eps": 1e-05, | |
| "max_position_embeddings": 514, | |
| "model_type": "bert", | |
| "num_attention_heads": 16, | |
| "num_hidden_layers": 6, | |
| "output_hidden_states": true, | |
| "pad_token_id": 1, | |
| "position_embedding_type": "relative_key", | |
| "torch_dtype": "float32", | |
| "transformers_version": "4.17.0", | |
| "type_vocab_size": 1, | |
| "use_cache": true, | |
| "vocab_size": 32000 | |
| } | |