Sergey Pinus

> Привет! Если бы вы могли поделиться, как вам удалось заставить его работать с RTX 3090, я был бы очень благодарен! У меня так много ошибок, что я не знаю,...

Не помогло. No result `◼️ Malevich is 1.3 billion params model from the family GPT3-like, that uses Russian language and text+image multi-modality. tokenizer --> ready Working with z of shape...

> > Не помогло. No result `◼️ > > 4 гб это по идее мало, если в модели 1.3 млрд параметров, даже если везде использовать fp16 её вес уже будет...

> добавлены примеры использования ruDALL-E Malevich (XL) с графическим процессором 3,5 Гб vRAM #51 > > [](https://colab.research.google.com/drive/1AoolDYePUpPkRCKIu0cP9zV7lX5QGD3Z?usp=sharing) > > или [версия Jupyter](https://github.com/sberbank-ai/ru-dalle/blob/master/jupyters/Malevich-3.5GB-vRAM-usage.ipynb): > > @bropines надеюсь, это поможет...

I idiot sorry)

> So ideally, we would be able to set not just speed, but speaking rate as in words per minute, but I don't know if that's possible I assumed that...

I realize it's outside the scope of the project, but if you happen to want it, I'll buy you lots of tea

Oh, but wait. https://github.com/coqui-ai/TTS/issues/1674 It's about getting srt from a file, and I need to voice from srt. Sorry, I'm stupid.

> There is a proposal to add a video coloring system. That is, the neural network together with the artist painted one frame together. Subsequent frames that have similar elements...

> The current method outputs something like > > out.mp4 > if interested how did you do that? automatically or by hand?