Question about Table 4 in the paper

[Kevinz-code]

Hi, Lin Chen and Jinsong Li

Thanks for your great work and code sharing.

After reading your paper, I have a little concern about the experiment settings in Table 4. Specifically, what is the SFT settings of the column `SFT with ShareGPT-4V'.

In the paper, it reads `Utilizing only our ShareGPT4V subset during the SFT stages resulted in a significant increase of 31.4 points in MME perception score'

My concern is: Does it mean replacing part of llava665k with sharegpt-4v-cap100k dataset during SFT stage? or Does it mean SFT the llava pretrained models with only the sharegpt-4v-cap100k?

Looking forward to your kindly reply!

Best,