| { | |
| "_name_or_path": "/cm/archive/namlh35/SATD/results/my-dataset-0.0.1/starcoder2-3b-multilabel_kfold_code-0.0.2/0", | |
| "architectures": [ | |
| "Starcoder2ForSequenceClassification" | |
| ], | |
| "attention_dropout": 0.1, | |
| "bos_token_id": 0, | |
| "embedding_dropout": 0.1, | |
| "eos_token_id": 0, | |
| "finetuning_task": "text-classification", | |
| "hidden_act": "gelu_pytorch_tanh", | |
| "hidden_size": 3072, | |
| "id2label": { | |
| "0": "LABEL_0", | |
| "1": "LABEL_1", | |
| "2": "LABEL_2", | |
| "3": "LABEL_3" | |
| }, | |
| "initializer_range": 0.018042, | |
| "intermediate_size": 12288, | |
| "label2id": { | |
| "0": 0, | |
| "1": 1 | |
| }, | |
| "max_position_embeddings": 16384, | |
| "max_query_length": 512, | |
| "max_seq_length": -4, | |
| "mlp_type": "default", | |
| "model_type": "starcoder2", | |
| "norm_epsilon": 1e-05, | |
| "norm_type": "layer_norm", | |
| "num_attention_heads": 24, | |
| "num_hidden_layers": 30, | |
| "num_key_value_heads": 2, | |
| "pad_token_id": 0, | |
| "problem_type": "multi_label_classification", | |
| "residual_dropout": 0.1, | |
| "rope_theta": 999999.4420358813, | |
| "sc_comment": false, | |
| "sliding_window": 4096, | |
| "torch_dtype": "float32", | |
| "transformers_version": "4.40.2", | |
| "use_bias": true, | |
| "use_cache": true, | |
| "vocab_size": 49152 | |
| } | |