yl4579/StyleTTS2

h100: Worse output & 20x slower inference?

Open

#89 geöffnet am 26. Nov. 2023

Auf GitHub ansehen
 (14 Kommentare) (0 Reaktionen) (0 zugewiesene Personen)Python (210 Forks)batch import
help wanted

Repository-Metriken

Stars
 (3.429 Stars)
PR-Merge-Metriken
 (Keine gemergten PRs in 30 T)

Beschreibung

We're testing finetuning on an h100 and 4090, here are the results:

4090: https://voca.ro/11mtxzLHzzih h100: https://voca.ro/15QldVjuG7nu

Almost identical finetune, but h100 is output is SIGNIFICANTLY worse. It isn't a config issue, and we've replicated it twice with LJSpeech as well.

4090 is also faster during training and considerably faster during inference, almost 20x faster than h100:

Screenshot_2023-11-26_at_5 01 16_PM

h100:

Screenshot_2023-11-24_at_3 46 08_PM

And during training, one epoch took the 4090 about 3 minutes, while the h100 took 4.12 minutes.

Does anyone know what could be going on here? Never seen an issue like this on an h100 before with a diffusion like model. Thanks

Contributor Guide