config.json

{
 "_name_or_path": "AstraMindAI/xtts2-gpt",
  "architectures": [
    "XttsGPT"
  ],
  "auto_map": {
    "AutoConfig": "AstraMindAI/xtts2-gpt--gpt_config.XTTSGPTConfig",
    "AutoModelForCausalLM": "AstraMindAI/xtts2-gpt--xtts2_gpt_modeling.XttsGPT"
  },
  "audio_config": {
    "fmax": 8000,
    "fmin": 0,
    "hop_length": 256,
    "mel_channels": 80,
    "mel_norms_file": null,
    "n_fft": 1024,
    "output_sample_rate": 24000,
    "power": 1.0,
    "sample_rate": 22050,
    "win_length": 1024
  },
  "char_limits": {
    "ar": 166,
    "cs": 186,
    "de": 253,
    "en": 250,
    "es": 239,
    "fr": 273,
    "hu": 224,
    "it": 213,
    "ja": 71,
    "ko": 95,
    "nl": 251,
    "pl": 224,
    "pt": 203,
    "ru": 182,
    "tr": 226,
    "zh": 82
  },
  "duration_const": 102400,
  "enable_redaction": false,
  "gpt_batch_size": 1,
  "gpt_checkpointing": false,
  "gpt_code_stride_len": 1024,
  "gpt_cond_chunk_len": 4,
  "gpt_cond_len": 30,
  "gpt_layers": 30,
  "gpt_max_audio_tokens": 605,
  "gpt_max_prompt_tokens": 70,
  "gpt_max_text_tokens": 402,
  "gpt_n_heads": 16,
  "gpt_n_model_channels": 1024,
  "gpt_num_audio_tokens": 1026,
  "gpt_number_text_tokens": 6681,
  "gpt_start_audio_token": 1024,
  "gpt_start_text_token": null,
  "gpt_stop_audio_token": 1025,
  "gpt_stop_text_token": null,
  "gpt_train_solo_embeddings": false,
  "gpt_use_masking_gt_prompt_approach": true,
  "gpt_use_perceiver_resampler": true,
  "kv_cache": true,
  "label_smoothing": 0.0,
  "languages": [
    "en",
    "es",
    "fr",
    "de",
    "it",
    "pt",
    "pl",
    "tr",
    "ru",
    "nl",
    "cs",
    "ar",
    "zh-cn",
    "hu",
    "ko",
    "ja",
    "hi"
  ],
  "max_ref_len": 30,
  "model_type": "xtts_gpt",
  "num_chars": 255,
  "perceiver_cond_length_compression": 256,
  "repetition_penalty": 5.0,
  "sound_norm_refs": false,
  "temperature": 0.75,
  "top_p": 0.85,
  "transformers_version": "4.45.1",
  "vocab_size": 256,
  "cond_d_vector_in_each_upsampling_layer": true,
  "d_vector_dim": 512,
  "decoder_input_dim": 1024,
  "input_sample_rate": 22050,
  "hifi_model_type": "xtts_hifigan",
  "output_hop_length": 256,
  "output_sample_rate": 24000,
  "resblock_dilation_sizes": [
    [
      1,
      3,
      5
    ],
    [
      1,
      3,
      5
    ],
    [
      1,
      3,
      5
    ]
  ],
  "resblock_kernel_sizes": [
    3,
    7,
    11
  ],
  "speaker_encoder_config": {
    "model_config": null,
    "model_name": "speaker_encoder",
    "preprocess_config": null,
    "speaker_embedding_dim": 512,
    "use_torch_spec": true
  },
  "upsample_initial_channel": 512,
  "upsample_kernel_sizes": [
    16,
    16,
    4,
    4
  ],
  "upsample_rates": [
    8,
    8,
    2,
    2
  ]
}