{
  "architectures": [
    "MCTCTModel"
  ],
  "attention_head_dim": 128,
  "attention_probs_dropout_prob": 0.3,
  "bos_token_id": 0,
  "conv_channels": null,
  "conv_dropout": 0.3,
  "conv_glu_dim": 1,
  "conv_kernel": [
    7
  ],
  "conv_stride": [
    3
  ],
  "ctc_loss_reduction": "sum",
  "ctc_zero_infinity": false,
  "eos_token_id": 2,
  "hidden_act": "relu",
  "hidden_dropout_prob": 0.3,
  "hidden_size": 512,
  "initializer_range": 0.02,
  "input_channels": 1,
  "input_feat_per_channel": 80,
  "intermediate_size": 20,
  "layer_norm_eps": 1e-05,
  "layerdrop": 0.3,
  "max_position_embeddings": 920,
  "model_type": "mctct",
  "num_attention_heads": 4,
  "num_conv_layers": 1,
  "num_hidden_layers": 4,
  "pad_token_id": 1,
  "torch_dtype": "float32",
  "transformers_version": "4.28.0.dev0",
  "vocab_size": 32
}