base/base.cfg

{
  "audio_encoder_config": {
    "n_mels": 80,
    "n_audio_ctx": 1500,
    "n_audio_state": 512,
    "n_audio_head": 8,
    "n_audio_layer": 6
  },
  "text_decoder_config": {
    "n_vocab": 51865,
    "n_text_ctx": 448,
    "n_text_state": 512,
    "n_text_head": 8,
    "n_text_layer": 6
  }
}