Can someone help me with AllTalk TTS

[F1re4]

Hi all. I would like to ask you this question: I tried to install it all day today but there was an error, I was able to launch the custom interface, but when I ginirazzi the following problem appears

ERROR: Exception in ASGI application
Traceback (most recent call last):
File "F:\textai\alltalk_tts-main\alltalk_environment\env\Lib\site-packages\starlette\responses.py", line 264, in call
await wrap(partial(self.listen_for_disconnect, receive))
File "F:\textai\alltalk_tts-main\alltalk_environment\env\Lib\site-packages\starlette\responses.py", line 260, in wrap
await func()
File "F:\textai\alltalk_tts-main\alltalk_environment\env\Lib\site-packages\starlette\responses.py", line 237, in listen_for_disconnect
message = await receive()
^^^^^^^^^^^^^^^
File "F:\textai\alltalk_tts-main\alltalk_environment\env\Lib\site-packages\uvicorn\protocols\http\h11_impl.py", line 535, in receive
await self.message_event.wait()
File "F:\textai\alltalk_tts-main\alltalk_environment\env\Lib\asyncio\locks.py", line 213, in wait
await fut
asyncio.exceptions.CancelledError: Cancelled by cancel scope 248286cee50

During handling of the above exception, another exception occurred:

+-+---------------- 1 ----------------
| Traceback (most recent call last):
| File "F:\textai\alltalk_tts-main\alltalk_environment\env\Lib\site-packages\starlette\responses.py", line 260, in wrap
| await func()
| File "F:\textai\alltalk_tts-main\alltalk_environment\env\Lib\site-packages\starlette\responses.py", line 249, in stream_response
| async for chunk in self.body_iterator:
| File "F:\textai\alltalk_tts-main\tts_server.py", line 506, in generate_audio_internal
| gpt_cond_latent, speaker_embedding = model.get_conditioning_latents(
| ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
| File "F:\textai\alltalk_tts-main\alltalk_environment\env\Lib\site-packages\torch\utils_contextlib.py", line 115, in decorate_context
| return func(*args, **kwargs)
| ^^^^^^^^^^^^^^^^^^^^^
| File "F:\textai\alltalk_tts-main\alltalk_environment\env\Lib\site-packages\TTS\tts\models\xtts.py", line 365, in get_conditioning_latents
| speaker_embedding = self.get_speaker_embedding(audio, load_sr)
| ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
| File "F:\textai\alltalk_tts-main\alltalk_environment\env\Lib\site-packages\torch\utils_contextlib.py", line 115, in decorate_context
| return func(*args, **kwargs)
| ^^^^^^^^^^^^^^^^^^^^^
| File "F:\textai\alltalk_tts-main\alltalk_environment\env\Lib\site-packages\TTS\tts\models\xtts.py", line 318, in get_speaker_embedding
| audio_16k = torchaudio.functional.resample(audio, sr, 16000)
| ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
| File "F:\textai\alltalk_tts-main\alltalk_environment\env\Lib\site-packages\torchaudio\functional\functional.py", line 1519, in resample
| kernel, width = _get_sinc_resample_kernel(
| ^^^^^^^^^^^^^^^^^^^^^^^^^^
| File "F:\textai\alltalk_tts-main\alltalk_environment\env\Lib\site-packages\torchaudio\functional\functional.py", line 1417, in _get_sinc_resample_kernel
| idx = torch.arange(-width, width + orig_freq, dtype=idx_dtype, device=device)[None, None] / orig_freq
| ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
| RuntimeError: CUDA error: operation not supported
| Compile with TORCH_USE_CUDA_DSA to enable device-side assertions.
|
+------------------------------------

If possible, how can you work with AllTalk TTS with an AMD RX6700xt GPU? I tried with ZLUDA but it didn’t work, or I configured it wrong. If this is still impossible, then there are some alternatives that would work with Koboldcpp for my video card, thanks in advance

[brianjorden]

Take a look at this: #377

[F1re4]

Take a look at this: #377

thanks, i fixed some by installing HIP SDK but get that one what should i so?
raise RuntimeError(f"File at path {self.path} does not exist.")
RuntimeError: File at path F:\textai\alltalk_tts-main\outputs\undefined does not exist.
when i use basic version and when i use AllTalk Generator
[AllTalk TTSGen] Hello, this is a preview of voice arnold.
An error occurred: CUDA error: invalid argument
Compile with TORCH_USE_CUDA_DSA to enable device-side assertions.
this one

[erew123]

@F1re4 This is everything I know about ROCm support:

• You will probably have to install Pytorch with ROCm... but I dont know, I dont have an AMD card to test with so I cannot test/debug anything personally.
• People have discussed AMD/ROCm support on this discussion forum post Basic AMD ROCm setup guide #132
• this is the last thing someone told me on the subject Note - AMD ROCm 6.2 performance issues #392 and I assume its this version https://pytorch.org/get-started/previous-versions/#linux-and-windows-9

Beyond that, I couldn't tell you anything