aieng-lab/t5-large_comment-type-python

1{2  "_name_or_path": "t5-large",3  "architectures": [4    "T5ForSequenceClassification"5  ],6  "classifier_dropout": 0.0,7  "d_ff": 4096,8  "d_kv": 64,9  "d_model": 1024,10  "decoder_start_token_id": 0,11  "dense_act_fn": "relu",12  "dropout_rate": 0.1,13  "eos_token_id": 1,14  "feed_forward_proj": "relu",15  "id2label": {16    "0": "LABEL_0",17    "1": "LABEL_1",18    "2": "LABEL_2",19    "3": "LABEL_3",20    "4": "LABEL_4"21  },22  "initializer_factor": 1.0,23  "is_encoder_decoder": true,24  "is_gated_act": false,25  "label2id": {26    "LABEL_0": 0,27    "LABEL_1": 1,28    "LABEL_2": 2,29    "LABEL_3": 3,30    "LABEL_4": 431  },32  "layer_norm_epsilon": 1e-06,33  "model_type": "t5",34  "n_positions": 512,35  "num_decoder_layers": 24,36  "num_heads": 16,37  "num_layers": 24,38  "output_past": true,39  "pad_token_id": 0,40  "problem_type": "multi_label_classification",41  "relative_attention_max_distance": 128,42  "relative_attention_num_buckets": 32,43  "task_specific_params": {44    "summarization": {45      "early_stopping": true,46      "length_penalty": 2.0,47      "max_length": 200,48      "min_length": 30,49      "no_repeat_ngram_size": 3,50      "num_beams": 4,51      "prefix": "summarize: "52    },53    "translation_en_to_de": {54      "early_stopping": true,55      "max_length": 300,56      "num_beams": 4,57      "prefix": "translate English to German: "58    },59    "translation_en_to_fr": {60      "early_stopping": true,61      "max_length": 300,62      "num_beams": 4,63      "prefix": "translate English to French: "64    },65    "translation_en_to_ro": {66      "early_stopping": true,67      "max_length": 300,68      "num_beams": 4,69      "prefix": "translate English to Romanian: "70    }71  },72  "torch_dtype": "bfloat16",73  "transformers_version": "4.48.3",74  "use_cache": true,75  "vocab_size": 3212876}77