Ecosyste.ms: Timeline
Browse the timeline of events for every public repo on GitHub. Data updated hourly from GH Archive.
GreenLandisaLie created a comment on an issue on SWivid/F5-TTS
Its working now both in chinese and english! Thanks! @SWivid Maybe its worth adding a 'ONNX' branch at [https://huggingface.co/SWivid/F5-TTS/tree/main](https://huggingface.co/SWivid/F5-TTS/tree/...
SWivid created a comment on an issue on SWivid/F5-TTS
nothing wrong with your setting i think. if the crash just have happened at the same training point, maybe we could see throught for data used at that specific point. for if you are using the a10...
SWivid created a comment on a pull request on SWivid/F5-TTS
@lpscr just temporarily lol. you can see from `dev` branch now, will go for that structure, though import and path dependencies not solved yet, will do that tomorrow when i wake up. after that fi...
mhenrichsen created a comment on an issue on SWivid/F5-TTS
@SWivid do you see anything wrong here? `accelerate launch finetune-cli.py --exp_name F5TTS_Base --learning_rate 0.00001 --batch_size_per_gpu 3500 --batch_size_type frame --max_samples 64 --grad_a...
SWivid created a comment on an issue on SWivid/F5-TTS
> @SWivid no. It's 600ish hours of 500k samples. Nothing longer than 25 seconds. It's weird. The training process with 38400 frames per gpu just goes well on A100 80G for us, and it usually use ...
CRCODE22 created a comment on an issue on SWivid/F5-TTS
Multi-style does achieve the same output would have been nice to have similar in podcasts-style tab tho since the add option is better than having for example 2-3 fixed speakers. If it just remains...
CRCODE22 closed an issue on SWivid/F5-TTS
An additional speaker for the podcast (total of 3 speakers), Add speed sliders to adjust voices in podcast as well for each speaker.
See history here I am opening it up again because it is not closed to my satisfaction and I do not find the issue closed a discussion and clarification should be allowed before closing: Previous...
CRCODE22 created a comment on an issue on SWivid/F5-TTS
Multi-style does achieve the same output would have been nice to have similar in podcasts-style tab tho since the add option is better than having for example 2-3 fixed speakers. If it just remains...
SWivid created a comment on an issue on SWivid/F5-TTS
https://github.com/SWivid/F5-TTS/blob/32c3ee77017d728321ffaa5d10e9f9f6cd44f20d/gradio_app.py#L48 Could temporarily change `str(cached_path("hf://SWivid/F5-TTS/F5TTS_Base/model_1200000.safetensors"...
mhenrichsen created a comment on an issue on SWivid/F5-TTS
@SWivid no. It's 600ish hours of 500k samples. Nothing longer than 25 seconds.
SWivid created a comment on an issue on SWivid/F5-TTS
How is the dataset constructed. Does it contain extremely few long samples?
mhenrichsen opened an issue on SWivid/F5-TTS
VRAM usage randomly spikes. Continues fine after restart.
Hi, I'm training on a single A100. I've tried multiple batch sizes, but the same thing keeps happening. I usually use between 10-25% of all available GPU VRAM, and then all of the sudden it ...
danielw97 opened an issue on SWivid/F5-TTS
Ability to specify finetuned model in the Gradio app
Hello, I've come to this repo recently, so apologies if I've missed something here. Things sound great, although I'm currently doing some finetuning tests from the original f5tts checkpoint with ...
SWivid created a comment on an issue on SWivid/F5-TTS
> `<Speaker1>` instead of (Speaker1)? or `|<Speaker1>|`