My GPU is RTX4090, and most of the models loaded through this plugin cannot work on this GPU.
The console always prompts for low memory.
I found a model card for flan-t5-xxl on the Huggingface website.
There is a description of how to use semi precision, but I don't understand how to write Python code.
Can you add an option to use semi precision?
And thank you very much for your work!
My GPU is RTX4090, and most of the models loaded through this plugin cannot work on this GPU. The console always prompts for low memory.
I found a model card for flan-t5-xxl on the Huggingface website. There is a description of how to use semi precision, but I don't understand how to write Python code.
Can you add an option to use semi precision? And thank you very much for your work!