F5-TTS: Error in inference. Audio output with no content, all silence
warnings.warn(
You have passed task=transcribe, but also have set forced_decoder_ids to [[1, None], [2, 50360]] which creates a conflict. forced_decoder_ids will be ignored in favor of
About this issue
- Original URL
- State: closed
- Created 2 years ago
- Comments: 51 (17 by maintainers)
Commits related to this issue
- fix. add dtype check for asr pipeline addressing #356 — committed to SWivid/F5-TTS by SWivid 2 years ago
- fix. add dtype check for asr pipeline addressing #356 — committed to spygaurad/F5-TTS by SWivid 2 years ago
Good news š it finally worked for me ā¦
i used this āā¦and other edit suggestions mentioned here i will provide screenshot to all what i did on utils_infer.py file:
really not sure which one exactly that made it works , but it worth knowing that
Thank you so muchļ¼ this is effective, the sound is outputļ¼cli & gradioļ¼
Thanks! It works for me tooļ¼need to use threeļ¼
@hongyu2024 so you have already tried
dtype=torch.float32?if working is fine, just ignore the warning
no install
TL;DR: If youāre running it on Windows in the UI, always provide a reference text in the Advanced Settings, to make it run.
OK. I found the issue. It lies in the transcription.
In this line: https://github.com/SWivid/F5-TTS/blob/b0f482421b03e187ee7ca1893458f383e2c289d3/src/f5_tts/infer/utils_infer.py#L126
I changed whisper-large-v3-turbo to whisper-base, and I was able to get the result after a while. I suspect that when it was whisper-large-v3-turbo, it just took way longer because itās a larger model. So when we see it āstuckā, itās actually running the transcription. It just was taking too much time.
Somehow when running in cli this isnāt an issue, but when running in gradio itās extremely slow.
Even when running the whisper-base, itās still very slow, compared to what it should be (on a 4090 GPU)
Potentially some conflicts exist between gradio and the huggingface pipeline libraries on Windows? Windows is a weird system.