Train Extraction failure on AMD GPU

I have an 7900xtx

In the training tab when I click to extract features, this is what happens to all the sliced audio samples

Starting pitch extraction with 16 cores on cuda:0 using rmvpe...
C:\Users\linsi\Downloads\Applio-3.2.9\env\lib\site-packages\librosa\util\files.py:10: UserWarning: pkg_resources is deprecated as an API. See https://setuptools.pypa.io/en/latest/pkg_resources.html. The pkg_resources package is slated for removal as early as 2025-11-30. Refrain from using this package or pin to Setuptools<81.
from pkg_resources import resource_filename
0%| | 0/1324 [00:00<?, ?it/s]An error occurred extracting file C:\Users\linsi\Downloads\Applio-3.2.9\logs\GI Juh\sliced_audios_16k\0_0_100.wav on cuda:0: CUDA error: invalid argument
Compile with TORCH_USE_CUDA_DSA to enable device-side assertions.

An error occurred extracting file C:\Users\linsi\Downloads\Applio-3.2.9\logs\GI Juh\sliced_audios_16k\0_0_10.wav on cuda:0: CUDA error: invalid argument
Compile with TORCH_USE_CUDA_DSA to enable device-side assertions.

Start from the training tab's feature-extraction entry point and reproduce the RMVPE extraction on the reported AMD 7900 XTX setup. Investigate the CUDA invalid-argument errors affecting the sliced audio samples; done means feature extraction completes successfully for those samples without the reported error.