search

zqhang / AnomalyCLIP

Official implementation for AnomalyCLIP (ICLR 2024)
186 stars 20 forks source link

Question about the reimplementation of the result of original CLIP #25

Open AllenYkl opened 2 weeks ago

AllenYkl commented 2 weeks ago

Hi there, Congrats for the great work!

In table 1, I ve noticed you also include the result of Original CLIP model Screenshot 2024-06-20 at 08 55 23

Could you please share the setting of this experiments? Cuz my reimplementation based on your code shows lots of differences than yours. Like

  1. the size of CLIP (Base large huge?)
  2. The fixed text prompts you used. ( The "encode_text_with_prompt_ensemble" method you implemented?)
  3. any modifications on the vision side ? like DPAM

Thank you so much for your patience!

zqhang commented 2 weeks ago

Hi, 1、We use the same large backbone as AnomalyCLIP. 2、For the original CLIP, we use two single prompts (refer to Appendix Baselines for details). 3、We do not modify the vision side.

AllenYkl commented 2 weeks ago

Hi, 1、We use the same large backbone as AnomalyCLIP. 2、For the original CLIP, we use two single prompts (refer to Appendix Baselines for details). 3、We do not modify the vision side.

Thx a lot!