honestabelink

Hello @onikienko , Sorry about this, the intent was _surprisingly_ to remove `'error'` and change `'exit'` to `'close'`. _surprisingly only in the sense I can't believe how much my brain...

@onikienko Yes but, sorry more _relevant information_ left out I've added additional functionality to extract files over stdout and handle the binary data. _unpublished on my branch_ When handling large...

For me, `pip install protobuf` did not add `protoc` to my path. @asa008, @YeRen123455 Verify your system can find `protoc`, by calling it in your command window. `protoc --help` If...

@voorjaar, When I try this I get Value: Error: [invokeFunction] unknown service 'quickInputService'. Do you remember the version of monaco-editor this worked for. Running inside electron if that makes a...

This sounds like what [15.ai](https://15.ai/guide) does. > Phonological conventions and ARPAbet You may notice that certain words are pronounced slightly differently than what you might expect. For example, suppose that...

I have a couple comments. If you are using a custom dataset. I know you mentioned you used ocotillo to label so... [Wav2Vec2Phoneme](https://huggingface.co/docs/transformers/model_doc/wav2vec2_phoneme) > - Wav2Vec2Phoneme uses the exact same...

Wav2Vec2-Large-LV60 finetuned for phonetic labels outputs https://huggingface.co/facebook/wav2vec2-lv-60-espeak-cv-ft >Wav2Vec2-Large-LV60 finetuned on multi-lingual Common Voice This checkpoint leverages the pretrained checkpoint [wav2vec2-large-lv60](https://huggingface.co/facebook/wav2vec2-large-lv60) and is fine-tuned on [CommonVoice](https://huggingface.co/datasets/common_voice) to recognize phonetic labels in...

@rkjoan Try reducing the batch size https://github.com/neonbjb/tortoise-tts/blob/52e912f5d1c013da6c99362bd50cb404ac80024f/tortoise/api.py#L168 Perhaps halving it. If you really want to get fancy you can break apart the CLVP computation into 2 steps. Just make sure...

Another point of failure can be from generating the `conditioning latents`. https://github.com/neonbjb/tortoise-tts/blob/52e912f5d1c013da6c99362bd50cb404ac80024f/tortoise/models/autoregressive.py#L390-L398 This is because for every sample is accumulated on the gpu, if you have many samples or longer...