fix(infer_v2): Correct the import path of BigVGAN's custom cuda kernel

This commit is contained in:
Yrom 2025-09-11 16:55:18 +08:00
parent 8336824c71
commit e409c4a19b
No known key found for this signature in database

View File

@ -100,12 +100,12 @@ class IndexTTS2:
if self.use_cuda_kernel:
# preload the CUDA kernel for BigVGAN
try:
from indextts.BigVGAN.alias_free_activation.cuda import load
from indextts.s2mel.modules.bigvgan.alias_free_activation.cuda import activation1d
anti_alias_activation_cuda = load.load()
print(">> Preload custom CUDA kernel for BigVGAN", anti_alias_activation_cuda)
except:
print(">> Preload custom CUDA kernel for BigVGAN", activation1d.anti_alias_activation_cuda)
except Exception as e:
print(">> Failed to load custom CUDA kernel for BigVGAN. Falling back to torch.")
print(f"{e!r}")
self.use_cuda_kernel = False
self.extract_features = SeamlessM4TFeatureExtractor.from_pretrained("facebook/w2v-bert-2.0")