| { | |
| "_name_or_path": "daekeun-ml/KoSimCSE-supervised-roberta-base", | |
| "architectures": [ | |
| "RobertaModel" | |
| ], | |
| "attention_probs_dropout_prob": 0.1, | |
| "base_model": "klue/roberta-base", | |
| "batch_size": 64, | |
| "bos_token_id": 0, | |
| "classifier_dropout": null, | |
| "dataset_dir": "/opt/ml/input/data/training/", | |
| "debug": "False", | |
| "device": "cuda:0", | |
| "eos_token_id": 2, | |
| "eval_steps": 50, | |
| "hidden_act": "gelu", | |
| "hidden_dropout_prob": 0.1, | |
| "hidden_size": 768, | |
| "initializer_range": 0.02, | |
| "intermediate_size": 3072, | |
| "layer_norm_eps": 1e-12, | |
| "local_rank": 0, | |
| "lr": 3e-05, | |
| "lr_scheduler_type": "linear", | |
| "max_position_embeddings": 512, | |
| "max_seq_len": 64, | |
| "model_type": "roberta", | |
| "num_attention_heads": 12, | |
| "num_epochs": 1, | |
| "num_hidden_layers": 12, | |
| "num_warmup_steps": 0, | |
| "output_dir": "/opt/ml/model/", | |
| "pad_token_id": 1, | |
| "position_embedding_type": "absolute", | |
| "rank": 0, | |
| "save_path": "../model", | |
| "seed": 42, | |
| "temperature": 0.05, | |
| "torch_dtype": "float32", | |
| "transformers_version": "4.40.2", | |
| "type_vocab_size": 2, | |
| "use_cache": true, | |
| "use_fp16": "False", | |
| "version": 1.0, | |
| "vocab_size": 50265, | |
| "world_size": 4 | |
| } | |