Lajavaness/bilingual-document-embedding
8225
1{2 "_name_or_path": "dangvantuan/bilingual_impl",3 "architectures": [4 "BilingualModel"5 ],6 "model_type": "bilingual",7 "auto_map": {8 "AutoConfig":"dangvantuan/bilingual_impl--config.BilingualConfig",9 "AutoModel": "dangvantuan/bilingual_impl--modeling.BilingualModel",10 "AutoModelForMaskedLM": "dangvantuan/bilingual_impl--modeling.BilingualForMaskedLM",11 "AutoModelForMultipleChoice": "dangvantuan/bilingual_impl--modeling.BilingualForMultipleChoice",12 "AutoModelForQuestionAnswering": "dangvantuan/bilingual_impl--modeling.BilingualForQuestionAnswering",13 "AutoModelForSequenceClassification": "dangvantuan/bilingual_impl--modeling.BilingualForSequenceClassification",14 "AutoModelForTokenClassification": "dangvantuan/bilingual_impl--modeling.BilingualForTokenClassification"15 },16 "attention_probs_dropout_prob": 0.1,17 "classifier_dropout": null,18 "bos_token_id": 0,19 "eos_token_id": 2,20 "hidden_act": "gelu",21 "hidden_dropout_prob": 0.1,22 "hidden_size": 1024,23 "initializer_range": 0.02,24 "intermediate_size": 4096,25 "layer_norm_eps": 1e-05,26 "max_position_embeddings": 8194,27 "num_attention_heads": 16,28 "num_hidden_layers": 24,29 "output_past": true,30 "pad_token_id": 1,31 "position_embedding_type": "absolute",32 "torch_dtype": "float16",33 "transformers_version": "4.38.2",34 "type_vocab_size": 1,35 "use_cache": true,36 "vocab_size": 25000237}