LEv145

Results 62 comments of LEv145

I think now the text on buttons should be displayed correctly Try the new version https://github.com/LEv145/--sd-webui-ar-plus/tree/3d534f36d5c63cdfbbd87179f174a6c9e7e4c8b2 UPD: https://github.com/LEv145/--sd-webui-ar-plus/tree/1da3bc6d2705488fd3766db4260c1300fed73530

It seems that the `interactive=False` parameter is responsible for this behavior of the code The strange thing is that this parameter is not in the code Hmm..

I tried to move away from inheritance Maybe it will work without a problem https://github.com/LEv145/--sd-webui-ar-plus/tree/LEv145-buttons-without-inheritance

Someone will need to test the buttons in auto1111 and forge, and then it will be possible to release it in main

It seems that only this needs to be fixed for full support: https://github.com/LEv145/--sd-webui-ar-plus/issues/28 But for the fix, I will need to install Forge and run debuggerπŸ‘€ On the other hand,...

Maybe this will fix the problem https://github.com/LEv145/--sd-webui-ar-plus/tree/f06348ad6ff5501556a52b3e229e0c363c4d65f4 ![ΠΈΠ·ΠΎΠ±Ρ€Π°ΠΆΠ΅Π½ΠΈΠ΅](https://github.com/user-attachments/assets/88b33a67-25d6-40e6-ad50-8aa853e71661)

Maybe problem with cpu https://github.com/ai-forever/ru-gpts/blob/master/examples/ruGPT3XL_generation.ipynb ![ΠΈΠ·ΠΎΠ±Ρ€Π°ΠΆΠ΅Π½ΠΈΠ΅](https://user-images.githubusercontent.com/40071054/218875610-117bb769-ddb8-4b1b-96d9-f70731ef2ba1.png)

Hi! Thanks for the idea! But I didn't fully understand it Can you give a visual example of your current nodes?

Если ΠΈΡΠΏΠΎΠ»ΡŒΠ·ΠΎΠ²Π°Ρ‚ΡŒ ΠΊΠΎΠ½ΠΊΡ€Π΅Ρ‚Π½Ρ‹Π΅ вСрсии ΠΈΠ· [ruGPT3XL_finetune_example.ipynb](https://github.com/ai-forever/ru-gpts/blob/master/examples/ruGPT3XL_finetune_example.ipynb), Ρ‚ΠΎ ошибка Ρ‚Π° ΠΆΠ΅: ```py [/usr/local/lib/python3.8/dist-packages/triton/code_gen.py](https://localhost:8080/#) in _compile(self, device, attributes, constants, num_warps, num_stages, force_nc_cache, *wargs, **meta) 548 tt_device = _triton.driver.cu_device(device.index, False) 549 # Compile...

> Hey @LEv145 , thank you for bringing that up! > > The `num_beams` parameter refers to the beam search decoding strategy for the model: see [HuggingFace explanation ](https://huggingface.co/blog/how-to-generate) Try...