config.json

{
  "_name_or_path": "/data/ruchao/workdir/SPAPL_KidsASR/egs/MyST/exp/whisper_small_en_trans_adapter_encdec_lr1e-4_bn32_zeroinit_2gpus_4ksteps/checkpoint-4000",
  "activation_dropout": 0.0,
  "activation_function": "gelu",
  "apply_spec_augment": false,
  "architectures": [
    "WhisperForConditionalGeneration"
  ],
  "attention_dropout": 0.0,
  "begin_suppress_tokens": [
    220,
    50256
  ],
  "bos_token_id": 50257,
  "classifier_proj_size": 256,
  "d_model": 768,
  "decoder_attention_heads": 12,
  "decoder_ffn_dim": 3072,
  "decoder_layerdrop": 0.0,
  "decoder_layers": 12,
  "decoder_start_token_id": 50257,
  "dropout": 0.0,
  "encoder_attention_heads": 12,
  "encoder_ffn_dim": 3072,
  "encoder_layerdrop": 0.0,
  "encoder_layers": 12,
  "eos_token_id": 50256,
  "forced_decoder_ids": [
    [
      1,
      50362
    ]
  ],
  "init_std": 0.02,
  "is_encoder_decoder": true,
  "mask_feature_length": 10,
  "mask_feature_min_masks": 0,
  "mask_feature_prob": 0.0,
  "mask_time_length": 10,
  "mask_time_min_masks": 2,
  "mask_time_prob": 0.05,
  "max_length": 448,
  "max_source_positions": 1500,
  "max_target_positions": 448,
  "median_filter_width": 7,
  "model_type": "whisper",
  "num_hidden_layers": 12,
  "num_mel_bins": 80,
  "pad_token_id": 50256,
  "peft_config": {
    "bottleneck_dim": 32,
    "dropout": 0.1,
    "lora_alpha": 128,
    "lora_dim": 8,
    "peft_decoder_layers": [
      1,
      2,
      3,
      4,
      5,
      6,
      7,
      8,
      9,
      10,
      11,
      12
    ],
    "peft_encoder_layers": [
      0,
      1,
      2,
      3,
      4,
      5,
      6,
      7,
      8,
      9,
      10,
      11,
      12
    ],
    "peft_type": "adapter",
    "prefix_dropout_rate": 0.0,
    "prefix_hidden_dim": 16,
    "prefix_n_layer": 12,
    "prefix_seq_len": [
      50,
      10
    ],
    "prompt_init_vocab": true,
    "prompt_n_tokens": [
      100,
      20
    ],
    "prompt_random_range": 0.5,
    "to_decoder": true,
    "to_encoder": true
  },
  "pif_loss_alpha": 1.0,
  "scale_embedding": false,
  "torch_dtype": "float32",
  "transformers_version": "4.36.2",
  "use_cache": false,
  "use_weighted_layer_sum": false,
  "vocab_size": 51864
}