xlnet-base-cased / config.json
xshubhamx's picture
Training in progress, epoch 1
9d653a7 verified
{
"_name_or_path": "xlnet/xlnet-base-cased",
"architectures": [
"XLNetForSequenceClassification"
],
"attn_type": "bi",
"bi_data": false,
"bos_token_id": 1,
"clamp_len": -1,
"d_head": 64,
"d_inner": 3072,
"d_model": 768,
"dropout": 0.1,
"end_n_top": 5,
"eos_token_id": 2,
"ff_activation": "gelu",
"id2label": {
"0": "Issue",
"1": "Court Discourse",
"2": "Conclusion",
"3": "Precedent Analysis",
"4": "Section Analysis",
"5": "Argument by Petitioner",
"6": "Fact",
"7": "Argument by Respondent",
"8": "Ratio",
"9": "Appellant",
"10": "Respondent",
"11": "Argument by Appellant",
"12": "Petitioner",
"13": "Judge",
"14": "Argument by Defendant"
},
"initializer_range": 0.02,
"label2id": {
"Appellant": 9,
"Argument by Appellant": 11,
"Argument by Defendant": 14,
"Argument by Petitioner": 5,
"Argument by Respondent": 7,
"Conclusion": 2,
"Court Discourse": 1,
"Fact": 6,
"Issue": 0,
"Judge": 13,
"Petitioner": 12,
"Precedent Analysis": 3,
"Ratio": 8,
"Respondent": 10,
"Section Analysis": 4
},
"layer_norm_eps": 1e-12,
"mem_len": null,
"model_type": "xlnet",
"n_head": 12,
"n_layer": 12,
"pad_token_id": 5,
"problem_type": "single_label_classification",
"reuse_len": null,
"same_length": false,
"start_n_top": 5,
"summary_activation": "tanh",
"summary_last_dropout": 0.1,
"summary_type": "last",
"summary_use_proj": true,
"task_specific_params": {
"text-generation": {
"do_sample": true,
"max_length": 250
}
},
"torch_dtype": "float32",
"transformers_version": "4.35.2",
"untie_r": true,
"use_mems_eval": true,
"use_mems_train": false,
"vocab_size": 32000
}