DAMO-NLP-SG/VCD

Issues

• About greedy search in VCD

  #18 opened 4 months ago by haohaodw
  1

• Discrepancy in VCD Reproduced Results for POPE Experiment

  #24 opened 2 months ago by jiyunBae007
  9

• Is the response of Qwen-VL very short?

  #25 opened 2 months ago by zifuwan
  2

• About seed

  #30 opened a month ago by alex1243423
  0

• ValueError when initing

  #28 opened a month ago by pspdada
  1

• dtype

  #27 opened a month ago by alex1243423
  1

• Ask for parameters

  #26 opened a month ago by maming109
  1

• Inquiry Regarding Discrepancy in Results Using the Same Model and Methodology as Presented in Your Paper

  #16 opened 5 months ago by obananas
  14

• Inference Sample

  #23 opened 3 months ago by Shinyzenith
  1

• Request for Parameter and Prompt Configurations Used in Table 3

  #20 opened 3 months ago by ShanWang-Shan
  3

• About the experiment results in the Table.1

  #21 opened 3 months ago by tbbbk
  2

• question about evaluation.

  #17 opened 4 months ago by tt6746690
  1

• unable to reproducing the results of llava

  #9 opened 6 months ago by frankRenlf
  10

• VCD in LVLMs

  #19 opened 4 months ago by sunzjz
  0

• About the Experimental Setup

  #15 opened 5 months ago by haohaodw
  4

• Issues about importing llava

  #14 opened 5 months ago by Melon-Xu
  0

• `model_kwargs_cd` is always a copy from `model_kwargs`, is it a bug?

  #13 opened 5 months ago by teamwong111
  11

• model_kwargs_cd and model_kwargs problem

  #7 opened 6 months ago by Mr-xiu
  3

• might be a waste of resources

  #12 opened 6 months ago by SDaoer
  1

• Multinomial Sampling May Not Ensure the Reliability of the Answer

  #10 opened 6 months ago by Eason-Qin
  1

• unable to reproduce the results of Table 7

  #11 opened 6 months ago by XueJiang16
  1

• Unable to reproduce results from greedy search

  #8 opened 6 months ago by qychen2001
  4

• If two confidence levels of original and distorted inputs are high and similar, can plausibility constraints make negative effect?

  #6 opened 9 months ago by QiushiYang
  0

• Special Hyperparameter value or Checkpoints are needed?

  #5 opened 9 months ago by yejipark-m
  2

• Decoding hyperparameter details

  #4 opened a year ago by yuezih
  6

• can you provide the image subsets you used for evaluation because the whole set of gqa and coco is so large

  #3 opened a year ago by yfzhang114
  1

• what about directly ablate the image or just use a BLACK image?

  #2 opened a year ago by DingchenYang99
  2

• probability tensor contains either `inf`, `nan` or element < 0

  #1 opened a year ago by zli999
  8