{
  "model": {
    "type": "streaming_enc_dec_asr",
    "bos_token_id": 1,
    "eos_token_id": 2,
    "pad_token_id": 0,
    "decoder_start_token_id": 1,
    "vocab_size": 32768,
    "sample_rate": 16000,
    "chunk_samples": 8000,
    "encoder": {
      "hidden_size": 620,
      "num_attention_heads": 8,
      "num_hidden_layers": 10,
      "head_size": 64
    },
    "decoder": {
      "hidden_size": 512,
      "num_attention_heads": 8,
      "num_hidden_layers": 10,
      "head_size": 64
    },
    "vad": {
      "filename": "silero_vad.onnx",
      "threshold": 0.5,
      "silence_duration_ms": 500,
      "prefix_padding_ms": 200
    },
    "moonshine": {
      "frontend_filename": "frontend.ort",
      "encoder_filename": "encoder.ort",
      "adapter_filename": "adapter.ort",
      "cross_kv_filename": "cross_kv.ort",
      "decoder_kv_filename": "decoder_kv.ort",
      "sample_buffer_size": 79,
      "conv1_buffer_size": 4,
      "conv2_buffer_size": 4,
      "total_lookahead": 16,
      "max_seq_len": 448,
      "tokens_per_second": 6.5,
      "seconds_per_memory_frame": 0.02,
      "max_segment_memory_frames": 500,
      "min_segment_memory_frames": 250,
      "left_context_frames": 160
    }
  },
  "search": {
    "diversity_penalty": 0.0,
    "do_sample": false,
    "max_length": 448,
    "min_length": 0,
    "num_beams": 1,
    "num_return_sequences": 1,
    "past_present_share_buffer": false,
    "repetition_penalty": 1.0,
    "temperature": 1.0,
    "top_k": 1,
    "top_p": 1.0
  }
}