kelvinxu/arctic-captions

Issues

• How to get start and How the convolution works out

  #39 opened 8 years ago by CathyW77
  3

• Step by step readme file

  #45 opened 5 years ago by Shanzaay
  0

• Instruction to visualize hard attention

  #27 opened 9 years ago by vyouman
  3

• question about grads of alphas in hard Attention

  #18 opened 9 years ago by denglixi
  4

• Post your evaluation score

  #20 opened 9 years ago by intuinno
  39

• 'sample = empty string' when trying to train flickr_30k

  #16 opened 9 years ago by noureldien
  6

• Datasets in .pkl format?

  #1 opened 10 years ago by jazzsaxmafia
  40

• Tensoflow implementation of this model?

  #44 opened 6 years ago by zBabar
  0

• about platform

  #17 opened 9 years ago by vincentliu9
  2

• Dataset Format and Flow of Code

  #43 opened 6 years ago by zBabar
  0

• Where can I get "model_name.npz" to run the Jupyter Notebook example?

  #42 opened 7 years ago by leqo-c
  1

• f_init = theano.function([ctx], [ctx]+init_state+init_memory, name='f_init', profile=False)

  #41 opened 7 years ago by shaoxuan92
  0

• Can I get pkl files?

  #40 opened 7 years ago by juyeonKang
  0

• Questions or bugs in the adam optimizer

  #36 opened 8 years ago by ysjakking
  1

• cannot figure out the code

  #30 opened 8 years ago by vanpersie32
  0

• MemoryError

  #24 opened 8 years ago by Lorne0
  1

• Trained Models

  #35 opened 8 years ago by ilkerkesen
  2

• no re-ranking?

  #11 opened 8 years ago by JonghwanMun
  2

• question about "doubly stochastic attention"

  #38 opened 8 years ago by zym1010
  2

• A bug in setting Adam optimizer learning rate?

  #34 opened 8 years ago by yuvval
  2

• why don't finetune cnn?

  #37 opened 8 years ago by vanpersie32
  1

• MemoryError in sampling process (gen_model in generate_caps.py)

  #12 opened 9 years ago
  4

• where is model_name.npz?

  #32 opened 8 years ago by fancyerii
  2

• Bug in doubly stochastic attention?

  #33 opened 8 years ago by ramavedantam
  0

• How soon does the sanity check look like working fine?

  #31 opened 8 years ago by Kyubyong
  0

• Split problem

  #29 opened 9 years ago by wenhuchen
  0

• argument 'model' in the 'main function' (generate_caps.py)

  #28 opened 9 years ago by Avi-123
  0

• Instructions to re-train the model: where to start?

  #26 opened 9 years ago by arvind2505
  0

• how to set decay_c when train the net?

  #25 opened 9 years ago by chenfsjz
  0

• train error!

  #23 opened 9 years ago by chenfsjz
  0

• Reproduce result

  #19 opened 9 years ago by WangYueFt
  5

• how to run and get_start?

  #22 opened 9 years ago by andyyuan78
  0

• Generate captions for custom images

  #21 opened 9 years ago by johnkorn
  0

• cannot replicate the result in the paper?

  #5 opened 10 years ago by zcyang
  7

• KeyError: 'A'

  #15 opened 9 years ago by evercherish
  12

• Training parameters for Flickr30k, coco dataset?

  #4 opened 10 years ago by hyunseokmin
  2

• metrics.py reference format?

  #13 opened 9 years ago by xlhdh
  5

• "NaN detected" when using 'attn_type': 'stochastic'

  #14 opened 9 years ago by quilby
  1

• Sample script to train the model

  #10 opened 9 years ago by zer0n
  1

• Question about sampling in stochastic attention

  #9 opened 9 years ago by jazzsaxmafia
  3

• Typo in the arXiv & ICML paper

  #8 opened 10 years ago by zhongwen
  1

• bias term has been added twice in LSTM layer

  #7 opened 10 years ago by zhongwen
  1

• Feature extraction

  #6 opened 10 years ago by hyunseokmin
  2

• MissingInputError from theano.function

  #3 opened 10 years ago by hyunseokmin
  6

• RuntimeWarning: numpy.ndarray size changed, may indicate binary incompatibility

  #2 opened 10 years ago by jnhwkim
  2