| { | |
| "_name_or_path": "/home/fschlatt/Models/set-encoder-base", | |
| "add_extra_token": true, | |
| "architectures": [ | |
| "SetEncoderElectraModel" | |
| ], | |
| "attention_probs_dropout_prob": 0.1, | |
| "backbone_model_type": "electra", | |
| "classifier_dropout": null, | |
| "depth": 100, | |
| "doc_length": 256, | |
| "embedding_size": 768, | |
| "hidden_act": "gelu", | |
| "hidden_dropout_prob": 0.1, | |
| "hidden_size": 768, | |
| "initializer_range": 0.02, | |
| "intermediate_size": 3072, | |
| "layer_norm_eps": 1e-12, | |
| "max_position_embeddings": 512, | |
| "model_type": "set-encoder", | |
| "num_attention_heads": 12, | |
| "num_hidden_layers": 12, | |
| "pad_token_id": 0, | |
| "pooling_strategy": "first", | |
| "position_embedding_type": "absolute", | |
| "query_length": 32, | |
| "sample_missing_docs": true, | |
| "summary_activation": "gelu", | |
| "summary_last_dropout": 0.1, | |
| "summary_type": "first", | |
| "summary_use_proj": true, | |
| "torch_dtype": "float32", | |
| "transformers_version": "4.41.2", | |
| "type_vocab_size": 2, | |
| "use_cache": true, | |
| "vocab_size": 30528 | |
| } | |