File size: 1,630 Bytes
bf0a145 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 | {
"format": "soulx-singer-mlx",
"source": "Soul-AILab/SoulX-Singer",
"official_code": "https://github.com/Soul-AILab/SoulX-Singer",
"precision": "4bit",
"dtype": "bfloat16",
"components": [
{
"component": "svs",
"source_file": "model.pt",
"tensor_count": 587,
"skipped": [],
"numel": 704471426,
"weight_files": [
"model-00001-of-00003.safetensors",
"model-00002-of-00003.safetensors",
"model-00003-of-00003.safetensors"
]
},
{
"component": "svc",
"source_file": "model-svc.pt",
"tensor_count": 544,
"skipped": [],
"numel": 698444162,
"weight_files": [
"model-00001-of-00003.safetensors",
"model-00002-of-00003.safetensors",
"model-00003-of-00003.safetensors"
]
}
],
"runtime": {
"svs": "SoulXSinger checkpoint, saved as MLX-compatible safetensors",
"svc": "SoulXSingerSVC checkpoint, saved as MLX-compatible safetensors",
"decoder": "full audio generation currently uses the official PyTorch bridge"
},
"quantization": {
"bits": 4,
"group_size": 64,
"mode": "affine",
"layout": "each quantized tensor is stored as .qweight/.scales/.biases"
},
"quantization_summary": [
{
"component": "svs",
"quantized_tensors": 289,
"kept_tensors": 298
},
{
"component": "svc",
"quantized_tensors": 270,
"kept_tensors": 274
}
],
"target_repo": "mlx-community/SoulX-Singer-4bit",
"collection": "https://huggingface.co/collections/mlx-community/soulx-singer-mlx-6a1c0525d0911ea400102840"
}
|