MCPcopy
hub / github.com/KittenML/KittenTTS / download_from_huggingface

Function download_from_huggingface

kittentts/get_model.py:58–98  ·  view source on GitHub ↗

Download model files from Hugging Face repository. Args: repo_id: Hugging Face repository ID cache_dir: Directory to cache downloaded files Returns: KittenTTS_1_Onnx: Instantiated model ready for use

(repo_id="KittenML/kitten-tts-nano-0.1", cache_dir=None)

Source from the content-addressed store, hash-verified

56
57
58def download_from_huggingface(repo_id="KittenML/kitten-tts-nano-0.1", cache_dir=None):
59 """Download model files from Hugging Face repository.
60
61 Args:
62 repo_id: Hugging Face repository ID
63 cache_dir: Directory to cache downloaded files
64
65 Returns:
66 KittenTTS_1_Onnx: Instantiated model ready for use
67 """
68 # Download config file first
69 config_path = hf_hub_download(
70 repo_id=repo_id,
71 filename="config.json",
72 cache_dir=cache_dir
73 )
74
75 # Load config
76 with open(config_path, 'r') as f:
77 config = json.load(f)
78
79 if config.get("type") not in ["ONNX1", "ONNX2"]:
80 raise ValueError("Unsupported model type.")
81
82 # Download model and voices files based on config
83 model_path = hf_hub_download(
84 repo_id=repo_id,
85 filename=config["model_file"],
86 cache_dir=cache_dir
87 )
88
89 voices_path = hf_hub_download(
90 repo_id=repo_id,
91 filename=config["voices"],
92 cache_dir=cache_dir
93 )
94
95 # Instantiate and return model
96 model = KittenTTS_1_Onnx(model_path=model_path, voices_path=voices_path, speed_priors=config.get("speed_priors", {}) , voice_aliases=config.get("voice_aliases", {}))
97
98 return model
99
100
101def get_model(repo_id="KittenML/kitten-tts-nano-0.1", cache_dir=None):

Callers 1

__init__Method · 0.85

Calls 1

KittenTTS_1_OnnxClass · 0.85

Tested by

no test coverage detected