FreedomIntelligence
/

llava-v1.5-7b-TRIM

Model card Files Files and versions Community

songdj commited on 25 days ago

Commit

d0bb0bc

•

1 Parent(s): b4b9676

Update README.md

Files changed (1) hide show

README.md +2 -2

README.md CHANGED Viewed

@@ -15,11 +15,11 @@ Inspired by human attention patterns in Visual Question Answering (VQA) tasks, T
 The TRIM method has been extensively tested across 12 datasets, and the results demonstrate a significant reduction in computational overhead while maintaining a consistent level of performance.
 This research marks a critical stride in efficient MLLM development, promoting greater accessibility and sustainability of high-performing models.
- <img src="./main.pdf" width="600" alt="TRIM" align="center" />
 TRIM significantly streamlines the computational process, reducing the number of image tokens by approximately 79%, processing time by 67%, and memory usage by 30% relative to the baseline (LLaVA-1.5-7B).
-<img src="./fig1.pdf" width="300" alt="stat2"/>
 ## How to use?

 The TRIM method has been extensively tested across 12 datasets, and the results demonstrate a significant reduction in computational overhead while maintaining a consistent level of performance.
 This research marks a critical stride in efficient MLLM development, promoting greater accessibility and sustainability of high-performing models.
+ <img src="./main.jpg" width="600" alt="TRIM" align="center" />
 TRIM significantly streamlines the computational process, reducing the number of image tokens by approximately 79%, processing time by 67%, and memory usage by 30% relative to the baseline (LLaVA-1.5-7B).
+<img src="./fig1.jpg" width="300" alt="stat2"/>
 ## How to use?