basilkr/MAL_DICT_CHECK4

1{2  "_name_or_path": "basilkr/Malasar_48WER",3  "activation_dropout": 0.0,4  "activation_function": "gelu",5  "apply_spec_augment": false,6  "architectures": [7    "WhisperForConditionalGeneration"8  ],9  "attention_dropout": 0.0,10  "begin_suppress_tokens": [11    220,12    5025713  ],14  "bos_token_id": 50257,15  "classifier_proj_size": 256,16  "d_model": 1280,17  "decoder_attention_heads": 20,18  "decoder_ffn_dim": 5120,19  "decoder_layerdrop": 0.0,20  "decoder_layers": 32,21  "decoder_start_token_id": 50258,22  "dropout": 0.0,23  "encoder_attention_heads": 20,24  "encoder_ffn_dim": 5120,25  "encoder_layerdrop": 0.0,26  "encoder_layers": 32,27  "eos_token_id": 50257,28  "forced_decoder_ids": null,29  "init_std": 0.02,30  "is_encoder_decoder": true,31  "mask_feature_length": 10,32  "mask_feature_min_masks": 0,33  "mask_feature_prob": 0.0,34  "mask_time_length": 10,35  "mask_time_min_masks": 2,36  "mask_time_prob": 0.05,37  "max_length": 448,38  "max_source_positions": 1500,39  "max_target_positions": 448,40  "model_type": "whisper",41  "num_hidden_layers": 32,42  "num_mel_bins": 80,43  "pad_token_id": 50257,44  "scale_embedding": false,45  "suppress_tokens": [],46  "torch_dtype": "float32",47  "transformers_version": "4.28.0",48  "use_cache": true,49  "use_weighted_layer_sum": false,50  "vocab_size": 5186551}52