aieng-lab/t5-small_commit-intent
05
1{2 "_name_or_path": "t5-small",3 "architectures": [4 "T5ForSequenceClassification"5 ],6 "classifier_dropout": 0.0,7 "d_ff": 2048,8 "d_kv": 64,9 "d_model": 512,10 "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 },20 "initializer_factor": 1.0,21 "is_encoder_decoder": true,22 "is_gated_act": false,23 "label2id": {24 "LABEL_0": 0,25 "LABEL_1": 1,26 "LABEL_2": 227 },28 "layer_norm_epsilon": 1e-06,29 "model_type": "t5",30 "n_positions": 512,31 "num_decoder_layers": 6,32 "num_heads": 8,33 "num_layers": 6,34 "output_past": true,35 "pad_token_id": 0,36 "relative_attention_max_distance": 128,37 "relative_attention_num_buckets": 32,38 "task_specific_params": {39 "summarization": {40 "early_stopping": true,41 "length_penalty": 2.0,42 "max_length": 200,43 "min_length": 30,44 "no_repeat_ngram_size": 3,45 "num_beams": 4,46 "prefix": "summarize: "47 },48 "translation_en_to_de": {49 "early_stopping": true,50 "max_length": 300,51 "num_beams": 4,52 "prefix": "translate English to German: "53 },54 "translation_en_to_fr": {55 "early_stopping": true,56 "max_length": 300,57 "num_beams": 4,58 "prefix": "translate English to French: "59 },60 "translation_en_to_ro": {61 "early_stopping": true,62 "max_length": 300,63 "num_beams": 4,64 "prefix": "translate English to Romanian: "65 }66 },67 "torch_dtype": "bfloat16",68 "transformers_version": "4.48.3",69 "use_cache": true,70 "vocab_size": 3212871}72 