|
{ |
|
"activation_dropout": 0.1, |
|
"activation_function": "gelu", |
|
"add_cross_attention": true, |
|
"architectures": [ |
|
"ProphetNetForCausalLM" |
|
], |
|
"attention_dropout": 0.1, |
|
"bos_token_id": 1, |
|
"decoder_ffn_dim": 32, |
|
"decoder_start_token_id": 0, |
|
"disable_ngram_loss": false, |
|
"dropout": 0.1, |
|
"encoder_ffn_dim": 32, |
|
"eos_token_id": 2, |
|
"eps": 0.0, |
|
"hidden_size": 16, |
|
"init_std": 0.02, |
|
"is_decoder": true, |
|
"is_encoder_decoder": false, |
|
"max_position_embeddings": 30, |
|
"model_type": "prophetnet", |
|
"ngram": 2, |
|
"num_buckets": 32, |
|
"num_decoder_attention_heads": 4, |
|
"num_decoder_layers": 4, |
|
"num_encoder_attention_heads": 4, |
|
"num_encoder_layers": 4, |
|
"pad_token_id": 0, |
|
"relative_max_distance": 128, |
|
"torch_dtype": "float32", |
|
"transformers_version": "4.28.0.dev0", |
|
"use_cache": true, |
|
"vocab_size": 30522 |
|
} |
|
|