| { | |
| "architectures": [ | |
| "T5WithLMHeadModel" | |
| ], | |
| "d_ff": 2048, | |
| "d_kv": 64, | |
| "d_model": 512, | |
| "decoder_start_token_id": 0, | |
| "dropout_rate": 0.1, | |
| "finetuning_task": null, | |
| "initializer_factor": 1.0, | |
| "is_decoder": false, | |
| "layer_norm_epsilon": 1e-06, | |
| "n_positions": 512, | |
| "num_heads": 8, | |
| "num_labels": 2, | |
| "num_layers": 6, | |
| "output_attentions": false, | |
| "output_hidden_states": false, | |
| "output_past": true, | |
| "pruned_heads": {}, | |
| "relative_attention_num_buckets": 32, | |
| "task_specific_params": { | |
| "summarization": { | |
| "early_stopping": true, | |
| "length_penalty": 2.0, | |
| "max_length": 200, | |
| "min_length": 30, | |
| "no_repeat_ngram_size": 3, | |
| "num_beams": 4, | |
| "prefix": "summarize: " | |
| }, | |
| "translation_en_to_de": { | |
| "early_stopping": true, | |
| "length_penalty": 2.0, | |
| "max_length": 300, | |
| "num_beams": 4, | |
| "prefix": "translate English to German: " | |
| }, | |
| "translation_en_to_fr": { | |
| "early_stopping": true, | |
| "length_penalty": 2.0, | |
| "max_length": 300, | |
| "num_beams": 4, | |
| "prefix": "translate English to French: " | |
| }, | |
| "translation_en_to_ro": { | |
| "early_stopping": true, | |
| "length_penalty": 2.0, | |
| "max_length": 300, | |
| "num_beams": 4, | |
| "prefix": "translate English to Romanian: " | |
| } | |
| }, | |
| "torchscript": false, | |
| "use_bfloat16": false, | |
| "vocab_size": 32128 | |
| } |