Could you please add the Q3? Cause 12GB VRAM doesn't seem to be enough :(
#2
by fawogin598 - opened
Thank you in advance :)
I don't recommend using Q3_K_M, but it will work.
https://huggingface.co/hanzogak/Wan2.1-Anisora-14B-I2V-480P-GGUF/blob/main/Wan2_1-Anisora-I2V-480P-14B_Q3_K_M.gguf
但为什么呢？我的3080 10G 也可以跑Q5模型，甚至Q8，虽然很慢
I don't recommend using Q3_K_M, but it will work.
https://huggingface.co/hanzogak/Wan2.1-Anisora-14B-I2V-480P-GGUF/blob/main/Wan2_1-Anisora-I2V-480P-14B_Q3_K_M.gguf
有什么差距呢
You should be using sageattention to make this chug along quicker. By the way any chance of Q6? I think for the quality boost at least
For me, on 12GB, this raises out of memory:
但为什么呢？我的3080 10G 也可以跑Q5模型，甚至Q8，虽然很慢
Could you please share your workflow? Or the model part of it? @AC116