search

42Shawn / LLaVA-PruMerge

LLaVA-PruMerge: Adaptive Token Reduction for Efficient Large Multimodal Models
Apache License 2.0
68 stars 4 forks source link

FLOPs #4

Open cjxcn opened 3 months ago

cjxcn commented 3 months ago

What methodology do you employ to evaluate the FLOPs of the models illustrated in the succeeding visual representation?

image

42Shawn commented 3 months ago

I have a paper, LLM Inference Unveiled: Survey and Roofline Model Insights. In this paper, we have developed and open-sourced an LLM-Viewer, which can do this kind of analysis.

dargma commented 2 months ago

Hi, I want to reproduce the data on the above [table 2] using LLM-Viewer. Could you share config files(*) which were used to run LLM-Viewer targeting LLaVA-1.5 w/ PruMerge?

wangxidong06 commented 1 month ago

+1

wangxidong06 commented 1 month ago

How can I reproduce these results ?