model_summary.json

{
  "model_name": "Inflect-Nano-v1",
  "language": "English",
  "speaker": "mark",
  "sample_rate": 24000,
  "n_mels": 80,
  "acoustic_params": 3465125,
  "vocoder_generator_params": 1167077,
  "total_inference_params": 4632202,
  "acoustic_config": {
    "vocab_size": 256,
    "tone_size": 16,
    "lang_size": 4,
    "n_mels": 80,
    "hidden": 168,
    "encoder_layers": 5,
    "decoder_layers": 6,
    "decoder_ff_mult": 3,
    "kernel_size": 7,
    "speaker_count": 2,
    "speaker_dim": 64,
    "dropout": 0.08,
    "sample_rate": 24000,
    "max_frames": 1400,
    "postnet_scale": 0.1,
    "use_frame_pitch": true,
    "abs_frame_bins": 512
  },
  "vocoder_config": {
    "variant": "snake_v2mid",
    "sample_rate": 24000,
    "n_fft": 1024,
    "hop_size": 256,
    "win_size": 1024,
    "num_mels": 80,
    "fmin": 0.0,
    "fmax": 12000.0,
    "resblock": "1",
    "upsample_rates": [
      8,
      8,
      2,
      2
    ],
    "upsample_kernel_sizes": [
      16,
      16,
      4,
      4
    ],
    "upsample_initial_channel": 144,
    "resblock_kernel_sizes": [
      3,
      7,
      11
    ],
    "resblock_dilation_sizes": [
      [
        1,
        3,
        5
      ],
      [
        1,
        3,
        5
      ],
      [
        1,
        3,
        5
      ]
    ],
    "activation": "snake"
  },
  "acoustic_checkpoint": "weights/inflect_nano_v1_acoustic.pt",
  "vocoder_checkpoint": "weights/inflect_nano_v1_vocoder.pt"
}