This model still consumes a lot of memory, How can we quantify this model?
#37
by
chfm
- opened
I hope it can be a bit smaller, and the loading speed should be faster.
Hi, you can try using the block swap node from kijai.
Is it possible to run the model on 4gb vram? (3050) laptop.
Is it possible to run the model on 4gb vram? (3050) laptop.
Gguf Q3_k_m from Civit yes but there is only gguf for v1.
This looks awesome. What O see tehrough video generation itchanges the face of person is there a way to keep face more to original ine? Thx
v8 effect is so good, plz add v8 gguf .