GT-Vision-Lab/VQA_LSTM_CNN

Issues

• Fail to repeat the accuracy of the pretrained VGG model

  #27 opened 8 years ago by xhzhao
  2

• How to process the multiple choice answer

  #35 opened 6 years ago by WangWenshan
  0

• Abstract scene parameters num_ans and num_output

  #34 opened 6 years ago by sanjass
  0

• UNk Token

  #33 opened 7 years ago by samarIbrahem
  0

• Unsupported marker type 0xf0

  #30 opened 7 years ago by woshiacai
  1

• out of memory

  #31 opened 7 years ago by woshiacai
  0

• no clue

  #32 opened 7 years ago by jijibn
  1

• Number of pretrained image features not matching with number of images in COCO

  #29 opened 7 years ago by franroldans
  0

• Trained model gets low accuracy on VQA server

  #28 opened 8 years ago by idansc
  2

• would you tell me more about the parameter and dataset?

  #26 opened 8 years ago by SeekPoint
  0

• Providing feedback through correct answer

  #25 opened 8 years ago by goodrahstar
  0

• Might be remove the second term of output in LSTM

  #21 opened 9 years ago by haooooooqi
  0

• Some problems while implementing with tensorflow

  #20 opened 8 years ago by chingyaoc
  1

• Script to evaluate new image using model saved from VQA_LSTM_CNN

  #4 opened 9 years ago by kumarabhinavgupta
  4

• Issue while trying to run the evaluation script

  #24 opened 8 years ago by robertsatya
  2

• setting for abstract?

  #22 opened 8 years ago by shinandrew
  3

• libcudnn.so.4 not found even I had run ' luarocks install CuDNN'

  #9 opened 9 years ago by andyyuan78
  3

• require 'cunn' and 'cutorch' in CPU mode

  #19 opened 9 years ago by NightFury13
  0

• Number of training picture

  #17 opened 9 years ago by andrewliao11
  0

• How to cite the model?

  #16 opened 9 years ago by ili3p
  3

• Ideas for NLP pre-processing and feature engineering

  #14 opened 9 years ago by honnibal
  16

• run prepro_img.lua failed

  #11 opened 9 years ago by andyyuan78
  2

• Bugs in filtering and encoding questions in prepro.py

  #12 opened 9 years ago by satwikkottur
  1

• th train.lua -backend nn failed!

  #10 opened 9 years ago by andyyuan78
  1

• JPG is actually a PNG

  #8 opened 9 years ago by nhynes
  1

• VQA preprocessing on OpenEnded Questions?

  #6 opened 9 years ago by tribhuvanesh
  1

• VQA_LSTM_CNN going out of memory on Titan X with 12 GB RAM

  #1 opened 9 years ago by kumarabhinavgupta
  6