linzhiqiu/t2v_metrics

Issues

• 请问多张图片用文本提示进行打分的时候，是否有什么prompt技巧呢？

  #20 opened 2 months ago by Nieleilei
  1

• How can I evaluate images without them being in path format? What if my images are in PIL or Torch form?

  #17 opened 3 months ago by rainarit
  1

• About the probabilty of Yes

  #19 opened 2 months ago by daiyixiang666
  1

• Adding new VLM?

  #16 opened 4 months ago by ApocalypsezZ
  2

• Is the code of "Improving Image Generation using VQAScore" available?

  #14 opened 4 months ago by SydCS
  1

• Text Generation from CLIP-T5 Flan

  #11 opened 5 months ago by reachomk
  6

• Conflict with modules named dataset

  #13 opened 5 months ago by piyushnags
  2

• Adapting VQAScore to a new VLM

  #7 opened 6 months ago by alzaia
  10

• When the prompts about GenAI-Bench for T2I generation will released?

  #4 opened 6 months ago by qingtian5
  4

• How many GPU VRAM do I need if I use the clip-flant5-xl model？

  #5 opened 9 months ago by ywdong
  1

• Support for weights loading in lower precision

  #8 opened 9 months ago by alzaia
  1

• Do you consider "YES", "Yes" and "yes" all as possible answers?

  #9 opened 9 months ago by lijing1996
  3

• ModuleNotFoundError when importing t2v_metrics

  #6 opened 9 months ago by aldoz-mila
  2