| { | |
| "architectures": [ | |
| "RobertaSequenceClassificationMuxed" | |
| ], | |
| "attention_probs_dropout_prob": 0.1, | |
| "binary_hadamard_epsilon": 0, | |
| "bos_token_id": 0, | |
| "demuxing_variant": "index", | |
| "eos_token_id": 2, | |
| "gaussian_hadamard_norm": 20, | |
| "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, | |
| "learn_muxing": 0, | |
| "max_position_embeddings": 514, | |
| "model_type": "roberta", | |
| "muxing_variant": "gaussian_hadamard", | |
| "num_attention_heads": 12, | |
| "num_hidden_layers": 12, | |
| "num_instances": 10, | |
| "pad_token_id": 1, | |
| "position_embedding_type": "absolute", | |
| "retrieval_loss_coeff": 1, | |
| "retrieval_percentage": 1, | |
| "task_loss_coeff": 0, | |
| "transformers_version": "4.4.2", | |
| "type_vocab_size": 1, | |
| "use_cache": true, | |
| "vocab_size": 50265 | |
| } | |