| { | |
| "model_type": "flaubert", | |
| "amp": 1, | |
| "attention_dropout": 0.1, | |
| "bos_index": 0, | |
| "bptt": 512, | |
| "clip_grad_norm": 5, | |
| "dropout": 0.1, | |
| "emb_dim": 768, | |
| "encoder_only": true, | |
| "eos_index": 1, | |
| "fp16": true, | |
| "gelu_activation": true, | |
| "group_by_size": true, | |
| "id2lang": { | |
| "0": "fr" | |
| }, | |
| "lang2id": { | |
| "fr": 0 | |
| }, | |
| "langs": [ | |
| "fr" | |
| ], | |
| "layer_norm_eps": 1e-06, | |
| "lg_sampling_factor": -1, | |
| "lgs": "fr", | |
| "mask_index": 5, | |
| "max_batch_size": 0, | |
| "max_vocab": -1, | |
| "mlm_steps": [ | |
| [ | |
| "fr", | |
| null | |
| ] | |
| ], | |
| "n_heads": 12, | |
| "n_langs": 1, | |
| "n_layers": 12, | |
| "n_words": 48701, | |
| "pad_index": 2, | |
| "pre_norm": false, | |
| "sample_alpha": 0, | |
| "share_inout_emb": true, | |
| "tokens_per_batch": -1, | |
| "unk_index": 3, | |
| "use_apex": true, | |
| "use_lang_emb": true, | |
| "word_blank": 0, | |
| "word_dropout": 0, | |
| "word_keep": 0.1, | |
| "word_mask": 0.8, | |
| "word_mask_keep_rand": "0.8,0.1,0.1", | |
| "word_pred": 0.15, | |
| "word_rand": 0.1, | |
| "word_shuffle": 0 | |
| } | |