blackteay/GPT-SoVITS
1
0
Fork 0
mirror of https://github.com/RVC-Boss/GPT-SoVITS.git synced 2026-07-01 02:59:25 +08:00
Code Issues Packages Projects Releases Wiki Activity

.

Browse Source
This commit is contained in:
XXXXRT666 2025-10-24 04:18:18 +01:00
parent 3b9bd46b3e
commit f25fcfb0e4
3 changed files with 5 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

1
GPT_SoVITS/Accelerate/PyTorch/t2s_model_abc.py
View File

@ -698,6 +698,7 @@ class CUDAGraphCacheABC(ABC):
if self.is_applicable:
assert bsz <= self.max_batch_size
if self.graph_cache.get(bsz) is None:
self.graph_cache[bsz] = Queue()
self.create_graph_cache(bsz)
return self.graph_cache[bsz].get()
else:

5
README.md
View File

@ -59,6 +59,7 @@ Unseen speakers few-shot fine-tuning demo:
| RTX 4060 Ti | 0.028 | UNK | 28 | Flash Attn Varlen CUDAGraph |
| Apple M4 | 0.16 | 1363ms | 1 | MLX Varlen |
| Apple M4 | 0.06 | UNK | 40 | MLX Varlen |
| i7-12700K | 0.28 | UNK | 40 | Torch Varlen |
</div>
@ -76,8 +77,8 @@ For users in China, you can [Click Here to use AutoDL](https://www.codewithgpu.c
| :------------: | :-------------: | :-----------: |
| Python 3.10 | PyTorch 2.5.1 | CUDA 12.4 |
| Python 3.11 | PyTorch 2.5.1 | CUDA 12.4 |
| Python 3.11 | PyTorch 2.7.0 | CUDA 12.8 |
| Python 3.11 | PyTorch 2.8.0 | Apple Silicon |
| Python 3.11 | PyTorch 2.9.0 | CUDA 12.8 |
| Python 3.11 | PyTorch 2.9.0 | Apple Silicon |
| Python 3.10 | PyTorch 2.8.0 | CPU |
</div>

2
requirements.txt
View File

@ -31,7 +31,7 @@ ctranslate2>=4.0,<5
fastapi[standard]>=0.115.2
fast_langdetect>=0.3.1
funasr==1.0.27
gradio==5.25.0
gradio==5.49.0
librosa==0.10.2
matplotlib>=3.10
numpy<2.0