NamCyan/starcoder2-3b-technical-debt-code-tesoro

1{2  "_name_or_path": "/cm/archive/namlh35/SATD/results/my-dataset-0.0.1/starcoder2-3b-multilabel_kfold_code-0.0.2/0",3  "architectures": [4    "Starcoder2ForSequenceClassification"5  ],6  "attention_dropout": 0.1,7  "bos_token_id": 0,8  "embedding_dropout": 0.1,9  "eos_token_id": 0,10  "finetuning_task": "text-classification",11  "hidden_act": "gelu_pytorch_tanh",12  "hidden_size": 3072,13  "id2label": {14    "0": "LABEL_0",15    "1": "LABEL_1",16    "2": "LABEL_2",17    "3": "LABEL_3"18  },19  "initializer_range": 0.018042,20  "intermediate_size": 12288,21  "label2id": {22    "0": 0,23    "1": 124  },25  "max_position_embeddings": 16384,26  "max_query_length": 512,27  "max_seq_length": -4,28  "mlp_type": "default",29  "model_type": "starcoder2",30  "norm_epsilon": 1e-05,31  "norm_type": "layer_norm",32  "num_attention_heads": 24,33  "num_hidden_layers": 30,34  "num_key_value_heads": 2,35  "pad_token_id": 0,36  "problem_type": "multi_label_classification",37  "residual_dropout": 0.1,38  "rope_theta": 999999.4420358813,39  "sc_comment": false,40  "sliding_window": 4096,41  "torch_dtype": "float32",42  "transformers_version": "4.40.2",43  "use_bias": true,44  "use_cache": true,45  "vocab_size": 4915246}47