Text-to-Speech
Transformers
Safetensors
Zonos
zonos3_mini
feature-extraction
tts
speech-synthesis
multilingual
voice-cloning
zonos3
zyphra
custom_code
Instructions to use Zyphra/ZONOS3-Mini-hf with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use Zyphra/ZONOS3-Mini-hf with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-to-speech", model="Zyphra/ZONOS3-Mini-hf", trust_remote_code=True)# pip install -U transformers accelerate # Load model directly from transformers import AutoModel model = AutoModel.from_pretrained("Zyphra/ZONOS3-Mini-hf", trust_remote_code=True, device_map="auto") - Zonos
How to use Zyphra/ZONOS3-Mini-hf with Zonos:
# pip install git+https://github.com/Zyphra/Zonos.git import torchaudio from zonos.model import Zonos from zonos.conditioning import make_cond_dict model = Zonos.from_pretrained("Zyphra/ZONOS3-Mini-hf", device="cuda") wav, sr = torchaudio.load("speaker.wav") # 5-10s reference clip speaker = model.make_speaker_embedding(wav, sr) cond = make_cond_dict(text="Hello, world!", speaker=speaker, language="en-us") codes = model.generate(model.prepare_conditioning(cond)) audio = model.autoencoder.decode(codes)[0].cpu() torchaudio.save("sample.wav", audio, model.autoencoder.sampling_rate) - Notebooks
- Google Colab
- Kaggle
Download config.json from Zyphra/ZONOS3-Mini-hf: direct link, hf CLI and curl.
- Browser
- Download file 6.72 kB
-
https://huggingface.co/Zyphra/ZONOS3-Mini-hf/resolve/main/config.json
- Command line
-
hf download hf://Zyphra/ZONOS3-Mini-hf/config.json
-
curl -L -o config.json https://huggingface.co/Zyphra/ZONOS3-Mini-hf/resolve/main/config.json
6.72 kB
| { | |
| "architectures": [ | |
| "Zonos3MiniForTextToSpeech" | |
| ], | |
| "auto_map": { | |
| "AutoConfig": "configuration_zonos3.Zonos3MiniConfig", | |
| "AutoModel": "modeling_zonos3.Zonos3MiniForTextToSpeech", | |
| "AutoModelForTextToWaveform": "modeling_zonos3.Zonos3MiniForTextToSpeech" | |
| }, | |
| "codec": { | |
| "codebook_dim": 8, | |
| "codebook_size": 4096, | |
| "fir_kernel": 127, | |
| "head_depth": 8, | |
| "head_hidden": 2048, | |
| "head_hop": 184, | |
| "hop": 2940, | |
| "latent_dim": 1024, | |
| "n_ffts": [ | |
| 736, | |
| 2208 | |
| ], | |
| "resample_new": 735, | |
| "resample_orig": 736, | |
| "resample_width": 17, | |
| "residual_codebooks": 9, | |
| "semantic_depth": 4, | |
| "semantic_hidden": 2048, | |
| "semantic_width": 768, | |
| "stages": [ | |
| { | |
| "context_frames": 150, | |
| "d_model": 1280, | |
| "dim_feedforward": 5120, | |
| "input_dimension": 1024, | |
| "num_heads": 20, | |
| "num_layers": 2, | |
| "output_dimension": 1280 | |
| }, | |
| { | |
| "context_frames": 300, | |
| "d_model": 768, | |
| "dim_feedforward": 3072, | |
| "input_dimension": 640, | |
| "num_heads": 12, | |
| "num_layers": 2, | |
| "output_dimension": 768 | |
| }, | |
| { | |
| "context_frames": 480, | |
| "d_model": 768, | |
| "dim_feedforward": 3072, | |
| "input_dimension": 384, | |
| "num_heads": 12, | |
| "num_layers": 2, | |
| "output_dimension": 768 | |
| }, | |
| { | |
| "context_frames": 480, | |
| "d_model": 768, | |
| "dim_feedforward": 3072, | |
| "input_dimension": 384, | |
| "num_heads": 12, | |
| "num_layers": 2, | |
| "output_dimension": 768 | |
| }, | |
| { | |
| "context_frames": 480, | |
| "d_model": 768, | |
| "dim_feedforward": 3072, | |
| "input_dimension": 384, | |
| "num_heads": 12, | |
| "num_layers": 2, | |
| "output_dimension": null | |
| } | |
| ] | |
| }, | |
| "dtype": "bfloat16", | |
| "generation": { | |
| "decode_context_frames": 30, | |
| "headroom_db": 3.0, | |
| "languages": { | |
| "ar": null, | |
| "bg": null, | |
| "bn": null, | |
| "cmn": "zh", | |
| "cs": null, | |
| "da": null, | |
| "de": "de", | |
| "el": null, | |
| "en_gb": "en", | |
| "en_us": "en", | |
| "es": "es", | |
| "fi": null, | |
| "fr_fr": "fr", | |
| "gu": null, | |
| "he": null, | |
| "hi": null, | |
| "hu": null, | |
| "id": null, | |
| "it": "it", | |
| "ja": "ja", | |
| "ka": null, | |
| "kn": null, | |
| "ko": "ko", | |
| "ml": null, | |
| "mr": null, | |
| "ms": null, | |
| "nl": null, | |
| "no": null, | |
| "or": null, | |
| "pa": null, | |
| "pl": null, | |
| "pt_br": "pt", | |
| "ro": null, | |
| "ru": null, | |
| "sk": null, | |
| "sv": null, | |
| "ta": null, | |
| "te": null, | |
| "th": null, | |
| "tl": null, | |
| "tr": null, | |
| "uk": null, | |
| "ur": null, | |
| "vi": null | |
| }, | |
| "markup_tags": [ | |
| "laugh", | |
| "breath", | |
| "sigh", | |
| "cough", | |
| "panting", | |
| "shouting", | |
| "crying" | |
| ], | |
| "min_p": 0.0, | |
| "output_limit_dbfs": -1.0, | |
| "overlap_frames": 4, | |
| "repetition_codebooks": 6, | |
| "repetition_penalty": 2.4, | |
| "repetition_window": 96, | |
| "silence_frame": [ | |
| 1918, | |
| 746, | |
| 3544, | |
| 234, | |
| 1267, | |
| 2914, | |
| 3944, | |
| 2849, | |
| 326, | |
| 1337 | |
| ], | |
| "silence_prefill_frames": 3, | |
| "silence_prefill_max_frames": 30, | |
| "temperature": 0.75, | |
| "top_p": 0.69, | |
| "topk": 80 | |
| }, | |
| "lm": { | |
| "audio_pad_id": 4097, | |
| "codebook_size": 4096, | |
| "condition_bucket_counts": [ | |
| 10, | |
| 10, | |
| 30 | |
| ], | |
| "condition_edges": { | |
| "fmax": [ | |
| 8437.5, | |
| 10242.2, | |
| 11507.8, | |
| 13359.4, | |
| 15164.1, | |
| 15585.9, | |
| 15750.0, | |
| 17765.6, | |
| 19945.3 | |
| ], | |
| "loudness": [ | |
| -29.778, | |
| -26.947, | |
| -25.047, | |
| -23.3, | |
| -21.743, | |
| -20.386, | |
| -19.186, | |
| -17.851, | |
| -15.967 | |
| ], | |
| "speaking_rate": [ | |
| 6.546, | |
| 8.182, | |
| 9.273, | |
| 10.31, | |
| 11.054, | |
| 11.667, | |
| 12.193, | |
| 12.656, | |
| 13.083, | |
| 13.476, | |
| 13.846, | |
| 14.196, | |
| 14.53, | |
| 14.887, | |
| 15.193, | |
| 15.517, | |
| 15.865, | |
| 16.208, | |
| 16.579, | |
| 16.972, | |
| 17.411, | |
| 17.898, | |
| 18.434, | |
| 19.069, | |
| 19.818, | |
| 20.816, | |
| 22.286, | |
| 24.722, | |
| 29.25 | |
| ] | |
| }, | |
| "condition_names": [ | |
| "loudness", | |
| "fmax", | |
| "speaking_rate" | |
| ], | |
| "dim": 1024, | |
| "eoa_id": 4096, | |
| "ffn_hidden": 2048, | |
| "gdn_head_dim": 128, | |
| "gdn_heads": 4, | |
| "gdn_kernel": 4, | |
| "gdn_period": 4, | |
| "head_dim": 32, | |
| "loss_softcap": 15.0, | |
| "max_seqlen": 32768, | |
| "moe_hidden": 1536, | |
| "moe_routed_dim": 256, | |
| "moe_topk": 2, | |
| "n_codebooks": 10, | |
| "n_experts": 64, | |
| "n_heads": 16, | |
| "n_kv_heads": 4, | |
| "n_layers": 11, | |
| "norm_eps": 1e-05, | |
| "reference_max_frames": 900, | |
| "rope_theta": 10000.0, | |
| "speaker_embedding_dim": 192, | |
| "text_rows": 8197, | |
| "text_vocab": 8196 | |
| }, | |
| "model_type": "zonos3_mini", | |
| "sampling_rate": 44100, | |
| "speaker_encoder": { | |
| "bottleneck_dim": 128, | |
| "embed_dim": 192, | |
| "feat_dim": 80, | |
| "frame_length_ms": 25.0, | |
| "frame_shift_ms": 10.0, | |
| "hidden_dim": 64, | |
| "in_channels": 32, | |
| "max_seconds": 60.0, | |
| "num_experts": 4, | |
| "num_mel_bins": 80, | |
| "sample_rate": 16000, | |
| "stages": [ | |
| [ | |
| [ | |
| 3, | |
| 32, | |
| 1, | |
| 1, | |
| 4 | |
| ], | |
| [ | |
| 3, | |
| 32, | |
| 1, | |
| 1, | |
| 4 | |
| ] | |
| ], | |
| [ | |
| [ | |
| 3, | |
| 64, | |
| 2, | |
| 1, | |
| 2 | |
| ], | |
| [ | |
| 3, | |
| 64, | |
| 1, | |
| 2, | |
| 2 | |
| ], | |
| [ | |
| 3, | |
| 64, | |
| 1, | |
| 2, | |
| 2 | |
| ], | |
| [ | |
| 3, | |
| 64, | |
| 1, | |
| 2, | |
| 2 | |
| ] | |
| ], | |
| [ | |
| [ | |
| 3, | |
| 128, | |
| 2, | |
| 1, | |
| 2 | |
| ], | |
| [ | |
| 3, | |
| 128, | |
| 1, | |
| 2, | |
| 2 | |
| ], | |
| [ | |
| 3, | |
| 128, | |
| 1, | |
| 2, | |
| 2 | |
| ], | |
| [ | |
| 3, | |
| 128, | |
| 1, | |
| 2, | |
| 2 | |
| ], | |
| [ | |
| 3, | |
| 128, | |
| 1, | |
| 2, | |
| 2 | |
| ] | |
| ], | |
| [ | |
| [ | |
| 3, | |
| 256, | |
| 2, | |
| 1, | |
| 1 | |
| ], | |
| [ | |
| 2, | |
| 256, | |
| 1, | |
| 2, | |
| 1 | |
| ], | |
| [ | |
| 2, | |
| 256, | |
| 1, | |
| 2, | |
| 1 | |
| ] | |
| ] | |
| ], | |
| "waveform_scale": 32768.0 | |
| }, | |
| "transformers_version": "4.57.3" | |
| } | |