buxiangzhiren/DDCap

Issues

• 为什么我的一直卡在这里RANK and WORLD_SIZE in environ: 0/1

  #41 opened 5 months ago by zhangfujunaaa
  1

• 模型结果

  #40 opened 10 months ago by songbohui
  0

• Your notebook tried to allocate more memory than is available. It has restarted.

  #39 opened a year ago by Eyantra698Sumanto
  0

• network problems

  #38 opened a year ago by songbohui
  1

• GPU利用率很低

  #37 opened a year ago by Sparkle-Q
  0

• oscar_split_ViT-B_32_train_512.pkl文件的来源

  #27 opened 2 years ago by masaiaaaa
  12

• 更换数据集问题

  #3 opened 2 years ago by Leng-bingo
  4

• pkl文件相关问题

  #5 opened 2 years ago by Leng-bingo
  12

• Release of predicted caption length

  #36 opened a year ago by baaaad
  0

• error

  #35 opened a year ago by songbohui
  5

• Set batchsize

  #29 opened a year ago by songbohui
  9

• 关于第一轮结束之后，关于java的报错问题（用五个卡跑的，java问题报错后， watch -n 2 --color gpustat --c显示还有四个卡在跑）

  #34 opened a year ago by Markkk111
  10

• How many parameters of your models?

  #33 opened a year ago by ShyFoo
  2

• 请问如何进行测试

  #32 opened a year ago by asasdsadagweh
  3

• On length prediction accuracy results in your paper

  #31 opened a year ago by xipq
  5

• Any formula for q_pred function ?

  #25 opened a year ago by linmou
  1

• Beam search

  #30 opened a year ago by vinevix
  7

• Where is the code for the inference stage

  #28 opened 2 years ago by xidian2wuzhaoji
  2

• How can I get parameters for using ScstRewardCriterion()?

  #26 opened 2 years ago by vinevix
  1

• alpha_schedule问题

  #24 opened 2 years ago by Iriya99
  6

• Can I train this model on Flickr30k?

  #23 opened 2 years ago by vinevix
  1

• What is the effect of the prefix_length?

  #22 opened 2 years ago by ShyFoo
  2

• 请问如何使用 caption 测评代码？

  #21 opened 2 years ago by verigle
  1

• CosineLRScheduler 没有t_mul 参数

  #9 opened 2 years ago by verigle
  31

• Why is there no coco test dataset?

  #20 opened 2 years ago by ShyFoo
  0

• Using different loss

  #19 opened 2 years ago by vinevix
  1

• epoch 10开始loss都为NaN

  #16 opened 2 years ago by ontheway-arch
  11

• How to use one GPU for training.

  #18 opened 2 years ago by shenyiyangsyy
  1

• I trained the model throughout 30 epochs, how can I just test it predicting captiong for 1 image?

  #17 opened 2 years ago by vinevix
  1

• [Errno 13] Permission denied: '/results_diff'

  #15 opened 2 years ago by ontheway-arch
  10

• Is there a link to download oscar_split_ViT-B_32_train_512.pkl with wget?

  #14 opened 2 years ago by vinevix
  6

• Training executes first Epoch but then stop itself, how come?

  #13 opened 2 years ago by vinevix
  9

• What should the high confidence score of tokens be set to ?

  #12 opened 2 years ago by greens007
  2

• train_tclip.py文件问题

  #11 opened 2 years ago by Leng-bingo
  2

• batchsize问题

  #7 opened 2 years ago by Leng-bingo
  20

• tmp-results-{}.json 保存问题 写入文件与保存文件位置不一致

  #10 opened 2 years ago by verigle
  2

• environment.yml 文件是正确 的吗？可以重新导出环境的配置文件吗？

  #8 opened 2 years ago by verigle
  2

• batch

  #6 opened 2 years ago by Leng-bingo
  0

• 关于self.bos_embedding的一些疑问

  #4 opened 2 years ago by benboerbababoerben
  2

• 运行程序问题

  #2 opened 2 years ago by Leng-bingo
  4

• Suspected information leakage in training & val

  #1 opened 2 years ago by ChangyaoTian
  8