{ | |
"architectures": [ | |
"AutoModelForSeq2SeqLM" | |
], | |
"model_type": "fsmt", | |
"activation_dropout": 0.0, | |
"activation_function": "relu", | |
"attention_dropout": 0.0, | |
"d_model": 256, | |
"dropout": 0.3, | |
"init_std": 0.02, | |
"max_position_embeddings": 1024, | |
"num_hidden_layers": 2, | |
"src_vocab_size": 1000, | |
"tgt_vocab_size": 1000, | |
"langs": [ | |
"src", | |
"trg" | |
], | |
"encoder_attention_heads": 4, | |
"encoder_ffn_dim": 1024, | |
"encoder_layerdrop": 0, | |
"encoder_layers": 2, | |
"decoder_attention_heads": 8, | |
"decoder_ffn_dim": 1024, | |
"decoder_layerdrop": 0, | |
"decoder_layers": 4, | |
"bos_token_id": 0, | |
"pad_token_id": 1, | |
"eos_token_id": 2, | |
"unk_token_id": 3, | |
"is_encoder_decoder": true, | |
"scale_embedding": true, | |
"tie_word_embeddings": true, | |
"num_beams": 5, | |
"early_stopping": false, | |
"length_penalty": 1.0 | |
} |