Thanks for your excellent work! Will the randomly selected 5k fine-tuning data be released?

ys-zong / VLGuard

[ICML 2024] Safety Fine-Tuning at (Almost) No Cost: A Baseline for Vision Large Language Models.

https://ys-zong.github.io/VLGuard/

45 stars 2 forks source link

Thanks for your excellent work! Will the randomly selected 5k fine-tuning data be released? #4

Closed ThBird closed 2 months ago

ys-zong commented 2 months ago

Hi, I didn't save the randomly selected samples. It's literally random.sample(llava_data, 5000). Different random seeds may lead to slightly different performance but should be similar.

ThBird commented 2 months ago

OK,Thanks