search

daswer123 / xtts-finetune-webui

Slightly improved official version for finetune xtts
199 stars 71 forks source link

cublas64_12.dll is not found or cannot be loaded #66

Open ChairmanMeow999 opened 1 day ago

ChairmanMeow999 commented 1 day ago

Thought i had this fixed but i guess not. I was getting the same error outlined in this post https://github.com/daswer123/xtts-finetune-webui/issues/56

I can't seem to be able to use the "create dataset" button without getting an error.. I followed the solution that was given here and i thought it fixed my issue but in reality it just got me to the point where it was downloading the model and fooled me into thinking it was actually working. I was doing some looking around in the issues section and found someone referencing it here https://github.com/daswer123/xtts-finetune-webui/issues/53#issuecomment-2333939074

He says at the end of the post "if you receive this error: Library cublas64_12.dll is not found or cannot be loaded you may need to go a pip upgrade for gradio to make sure you now have the proper dll for pydantic/fastapi."

And well that's the exact error message im getting and i thought i followed everything he said but im still getting the issue... I've tried deleting everything ai related included cuda and python and reinstalling it and im still getting the same error i've even gone through everything in the post he referenced in his comment https://github.com/gradio-app/gradio/issues/9278

I've tried everything and still no dice... if anyone has any ideas that would be great. Here is my error message.

The data processing was interrupted due an error !! Please check the console to verify the full error message! Error summary: Traceback (most recent call last): File "D:\stable3\TTS\xtts-finetune-webui\xtts_demo.py", line 246, in preprocess_dataset train_meta, eval_meta, audio_total_size = format_audio_list(audio_path, whisper_model = whisper_model, target_language=language, out_path=out_path, gradio_progress=progress) File "D:\stable3\TTS\xtts-finetune-webui\utils\formatter.py", line 131, in format_audio_list segments = list(segments) File "D:\stable3\TTS\xtts-finetune-webui\venv\lib\site-packages\faster_whisper\transcribe.py", line 1111, in restore_speech_timestamps for segment in segments: File "D:\stable3\TTS\xtts-finetune-webui\venv\lib\site-packages\faster_whisper\transcribe.py", line 508, in generate_segments encoder_output = self.encode(segment) File "D:\stable3\TTS\xtts-finetune-webui\venv\lib\site-packages\faster_whisper\transcribe.py", line 767, in encode return self.model.encode(features, to_cpu=to_cpu) RuntimeError: Library cublas64_12.dll is not found or cannot be loaded

epicstar7 commented 7 hours ago

From https://github.com/daswer123/xtts-finetune-webui/issues/15#issuecomment-2110821552

go inside folder xtts-finetune-webui\venv\Lib\site-packages\torch\lib find files cublas64_11.dll and cublasLt64_11.dll copy them inside same folder rename them as cublas64_12.dll and cublasLt64_12.dll so you will have 4 files cublas64_11.dll cublas64_12.dll cublasLt64_11.dll cublasLt64_12.dll problem fixed

ChairmanMeow999 commented 1 hour ago

From #15 (comment)

go inside folder xtts-finetune-webui\venv\Lib\site-packages\torch\lib find files cublas64_11.dll and cublasLt64_11.dll copy them inside same folder rename them as cublas64_12.dll and cublasLt64_12.dll so you will have 4 files cublas64_11.dll cublas64_12.dll cublasLt64_11.dll cublasLt64_12.dll problem fixed

Ok that definitely helped i was able to actually click the create dataset button and it even got to 100% but then it gave me an blank error with nothing appearing in the console. although it did output files into "finetune_models" anyone know why that might be?