Open rhmaaa opened 2 months ago
@ZeyuMi @eltociear @MatthewCroughan @hodlen
There's a lot of missing information that makes reproducibility hard, and the code for the newer PowerInfer-2 is not yet open, I've found.
There's a lot of missing information that makes reproducibility hard, and the code for the newer PowerInfer-2 is not yet open, I've found.
Thank you! I was just wondering how you got the data in Figure 4 and how to modify dejavu?
Use cupy to load model ?
@MatthewCroughan
I want to reproduce the results of fig4 in your paper, but I have encountered many problems. Can you provide some ideas or codes?