Utilities>Tokenizer does not work after TTS is loaded #145

New Issue

psammites · 2023-03-17T09:26:18Z

psammites commented

2023-03-17 09:26:18 +00:00

===================================BUG REPORT===================================
Welcome to bitsandbytes. For bug reports, please submit your error trace to: https://github.com/TimDettmers/bitsandbytes/issues
================================================================================
Running on local URL:  http://172.20.8.87:8080

To create a public link, set `share=True` in `launch()`.
Loading TorToiSe... (AR: ./training/Slovenian/finetune/models/200_gpt.pth, vocoder: bigvgan_24khz_100band)
Hardware acceleration found: cuda
Loading tokenizer JSON: ./modules/tortoise-tts/tortoise/data/tokenizer.json
Loaded tokenizer
Loading autoregressive model: ./training/Slovenian/finetune/models/200_gpt.pth
Loaded autoregressive model
Loaded diffusion model
Loading vocoder model: bigvgan_24khz_100band
Loading vocoder model: bigvgan_24khz_100band.pth
Removing weight norm...
Loaded vocoder model
Loaded TorToiSe, ready for generation.
Traceback (most recent call last):
  File "/home/sneed/ai-voice-cloning/venv/lib/python3.10/site-packages/gradio/routes.py", line 384, in run_predict
    output = await app.get_blocks().process_api(
  File "/home/sneed/ai-voice-cloning/venv/lib/python3.10/site-packages/gradio/blocks.py", line 1032, in process_api
    result = await self.call_function(
  File "/home/sneed/ai-voice-cloning/venv/lib/python3.10/site-packages/gradio/blocks.py", line 844, in call_function
    prediction = await anyio.to_thread.run_sync(
  File "/home/sneed/ai-voice-cloning/venv/lib/python3.10/site-packages/anyio/to_thread.py", line 31, in run_sync
    return await get_asynclib().run_sync_in_worker_thread(
  File "/home/sneed/ai-voice-cloning/venv/lib/python3.10/site-packages/anyio/_backends/_asyncio.py", line 937, in run_sync_in_worker_thread
    return await future
  File "/home/sneed/ai-voice-cloning/venv/lib/python3.10/site-packages/anyio/_backends/_asyncio.py", line 867, in run
    result = context.run(func, *args)
  File "/home/sneed/ai-voice-cloning/src/utils.py", line 1899, in tokenize_text
    encoded = tokenizer.encode(text)
UnboundLocalError: local variable 'tokenizer' referenced before assignment

Edit: With "Do Not Load TTS On Startup" Utilities>Tokenizer will work until "(Re)Load TTS" is clicked or audio is generated, after that the same error occurs as if TTS was loaded on startup.

``` ===================================BUG REPORT=================================== Welcome to bitsandbytes. For bug reports, please submit your error trace to: https://github.com/TimDettmers/bitsandbytes/issues ================================================================================ Running on local URL: http://172.20.8.87:8080 To create a public link, set `share=True` in `launch()`. Loading TorToiSe... (AR: ./training/Slovenian/finetune/models/200_gpt.pth, vocoder: bigvgan_24khz_100band) Hardware acceleration found: cuda Loading tokenizer JSON: ./modules/tortoise-tts/tortoise/data/tokenizer.json Loaded tokenizer Loading autoregressive model: ./training/Slovenian/finetune/models/200_gpt.pth Loaded autoregressive model Loaded diffusion model Loading vocoder model: bigvgan_24khz_100band Loading vocoder model: bigvgan_24khz_100band.pth Removing weight norm... Loaded vocoder model Loaded TorToiSe, ready for generation. Traceback (most recent call last): File "/home/sneed/ai-voice-cloning/venv/lib/python3.10/site-packages/gradio/routes.py", line 384, in run_predict output = await app.get_blocks().process_api( File "/home/sneed/ai-voice-cloning/venv/lib/python3.10/site-packages/gradio/blocks.py", line 1032, in process_api result = await self.call_function( File "/home/sneed/ai-voice-cloning/venv/lib/python3.10/site-packages/gradio/blocks.py", line 844, in call_function prediction = await anyio.to_thread.run_sync( File "/home/sneed/ai-voice-cloning/venv/lib/python3.10/site-packages/anyio/to_thread.py", line 31, in run_sync return await get_asynclib().run_sync_in_worker_thread( File "/home/sneed/ai-voice-cloning/venv/lib/python3.10/site-packages/anyio/_backends/_asyncio.py", line 937, in run_sync_in_worker_thread return await future File "/home/sneed/ai-voice-cloning/venv/lib/python3.10/site-packages/anyio/_backends/_asyncio.py", line 867, in run result = context.run(func, *args) File "/home/sneed/ai-voice-cloning/src/utils.py", line 1899, in tokenize_text encoded = tokenizer.encode(text) UnboundLocalError: local variable 'tokenizer' referenced before assignment ``` Edit: With "Do Not Load TTS On Startup" Utilities>Tokenizer will work until "(Re)Load TTS" is clicked or audio is generated, after that the same error occurs as if TTS was loaded on startup.

Screenshot 2023-03-17 020859.png

42 KiB

Screenshot 2023-03-17 022821.png

121 KiB

psammites changed title from ~~Clicking "(Re)Load TTS" breaks Utils>Tokenizer~~ to Utilities>Tokenizer does not work after TTS is loaded

2023-03-17 09:34:26 +00:00

nk990 commented

2023-03-17 10:32:54 +00:00

I saw you started to train the Slovenian model, any results, could you share some examples?

Sorry it out the topic

I saw you started to train the Slovenian model, any results, could you share some examples? ----- Sorry it out the topic

Screenshot from 2023-03-17 11-28-02.png

9.4 KiB

mrq commented

2023-03-17 13:08:53 +00:00

Should be fixed in commit 96b7f9d2cc.

Should be fixed in commit 96b7f9d2cc857b72f4a7de5ebb1875b140d272f8.

mrq closed this issue

2023-03-17 13:08:53 +00:00

psammites commented

2023-03-17 14:38:09 +00:00

I saw you started to train the Slovenian model, any results, could you share some examples?

@nk990 As of 200 iterations it wasn't any good. I might have picked the wrong learning rates, I'll update #133 if I get anything usable.

> I saw you started to train the Slovenian model, any results, could you share some examples? @nk990 As of 200 iterations it wasn't any good. I might have picked the wrong learning rates, I'll update #133 if I get anything usable.

Sign in to join this conversation.