GitHub / sgrvinod/a-pytorch-tutorial-to-image-captioning issues and pull requests
#206 - BLEU scores
Issue -
State: open - Opened by ZhangYiyii 3 months ago
#205 - VALIDATION QUERY
Issue -
State: open - Opened by ali4321143 8 months ago
#204 - eval.py gives 0 BLEU 4 score
Issue -
State: open - Opened by BlackBeard009 8 months ago
#203 - Query
Issue -
State: open - Opened by malikfahadsarwar 8 months ago
#100 - RuntimeError: size mismatch, m1: [16 x 3], m2: [2048 x 512]
Issue -
State: closed - Opened by ghost over 5 years ago
#99 - Chinese image caption, In the result, multiple words of the same type appear
Issue -
State: open - Opened by cylvzj over 5 years ago
#98 - RuntimeError : CUDA error: out of memory
Issue -
State: open - Opened by Tongyuang over 5 years ago
- 8 comments
#97 - BLEU-4 score @ beam size of 5 is 0.0606. result on author's model.
Issue -
State: open - Opened by ada-dl over 5 years ago
- 8 comments
#96 - python3.6/site-packages/torch/serialization.py:292: UserWarning: Couldn't retrieve source code for
Issue -
State: closed - Opened by ada-dl over 5 years ago
- 2 comments
#95 - Data location and WORDMAP file issue
Issue -
State: closed - Opened by ada-dl over 5 years ago
- 2 comments
#94 - Question regarding removing <start> from the targets during training
Issue -
State: closed - Opened by nilinykh over 5 years ago
- 5 comments
#93 - added map_location=str(device) in torch.load()
Pull Request -
State: closed - Opened by ngshya over 5 years ago
- 2 comments
#92 - AttributeError: 'Meteor' object has no attribute 'lock'
Issue -
State: open - Opened by Sun-WeiZhen over 5 years ago
- 1 comment
#91 - init_embedding
Issue -
State: open - Opened by zplovekq over 5 years ago
- 1 comment
#90 - unable to reproduce given validation BLEU-4 score.
Issue -
State: closed - Opened by mayankkumar-materialplus over 5 years ago
- 2 comments
#89 - How to do ensemble at testing with beam search?
Issue -
State: open - Opened by homelifes over 5 years ago
- 1 comment
#88 - Run captioning to my own images.
Issue -
State: open - Opened by gxskar almost 6 years ago
- 5 comments
#87 - in train.py epochs = 120?
Issue -
State: closed - Opened by lzy119 almost 6 years ago
#86 - Running on latest PyTorch 1.2.0
Issue -
State: open - Opened by fikrisyad almost 6 years ago
- 16 comments
#85 - What are saved to "TRAIN_IMAGES_flickr8k_5_cap_per_img_5_min_word_freq.hdf5" file?
Issue -
State: open - Opened by phyukhaing7 almost 6 years ago
- 1 comment
#84 - beam search's prev_word_inds and next_word_inds
Issue -
State: closed - Opened by fikrisyad almost 6 years ago
- 5 comments
#83 - why don't remove <end> for References and Hypotheses in validate ?
Issue -
State: open - Opened by Lzhushuai almost 6 years ago
- 3 comments
#82 - Train on Windows
Issue -
State: open - Opened by michael20at about 6 years ago
- 1 comment
#81 - License
Issue -
State: closed - Opened by vijaydone about 6 years ago
- 3 comments
#80 - scipy.misc import imread, imresize removed 2019
Issue -
State: open - Opened by piperino11 about 6 years ago
- 5 comments
#79 - multi layer RNN
Pull Request -
State: open - Opened by WoodySG2018 about 6 years ago
- 4 comments
#78 - If the length of the correct description is not equal to the predicted description length, the bleu evaluation cannot be performed.
Issue -
State: open - Opened by hubin111 about 6 years ago
- 4 comments
#77 - can you offer a tutorial for video captioning with soft attention?
Issue -
State: open - Opened by tuyunbin about 6 years ago
#76 - Get error loading the model when switch resnet101 to resnet50
Issue -
State: open - Opened by Ivanclj about 6 years ago
#75 - valueerror: too many values to unpack from pack_padded_sequence
Issue -
State: closed - Opened by Ivanclj about 6 years ago
- 8 comments
#74 - How to add evaluation indicators in eval.py such as Meteor, Rouge-L and Cider.
Issue -
State: open - Opened by RunCode90 about 6 years ago
- 1 comment
#73 - Two optimizers?
Issue -
State: open - Opened by johntiger1 about 6 years ago
#72 - scores in Beam search should be multiplied or added?
Issue -
State: open - Opened by Zachary-ZS about 6 years ago
- 1 comment
#71 - do you need to put decoder.dropout at testing time?
Issue -
State: closed - Opened by homelifes over 6 years ago
- 3 comments
#70 - checkpoint
Issue -
State: open - Opened by lzy119 over 6 years ago
- 4 comments
#69 - Q:for t in range(max(decode_lengths)):
Issue -
State: closed - Opened by kangkang59812 over 6 years ago
#68 - what is 'gate' used for ?
Issue -
State: closed - Opened by kangkang59812 over 6 years ago
- 4 comments
#67 - skimage warning
Issue -
State: open - Opened by volquelme over 6 years ago
- 1 comment
#66 - Caption generation error
Issue -
State: open - Opened by nattari over 6 years ago
- 1 comment
#65 - Weird result with your upload model and word map
Issue -
State: open - Opened by guantinglin over 6 years ago
- 2 comments
#64 - ValueError when evaluation
Issue -
State: closed - Opened by guantinglin over 6 years ago
- 6 comments
#63 - Evaluation issue
Issue -
State: open - Opened by anand-kummari over 6 years ago
- 1 comment
#62 - More evaluation methods including CIDEr, METEOR and ROUGE besides BLEU1-4
Pull Request -
State: open - Opened by yolo615 over 6 years ago
- 6 comments
#61 - Question about the checkpoint
Issue -
State: open - Opened by Jazz1996 over 6 years ago
- 8 comments
#60 - Why do you divide the image tensor by 255.
Issue -
State: open - Opened by dwang68 over 6 years ago
- 2 comments
#59 - cpu 1700%
Issue -
State: open - Opened by kangkang59812 over 6 years ago
- 3 comments
#58 - update to solve issue #57
Pull Request -
State: closed - Opened by kmario23 over 6 years ago
- 1 comment
#57 - RuntimeError: CUDA error: out of memory
Issue -
State: closed - Opened by kmario23 over 6 years ago
- 3 comments
#56 - eval.py cannot run-- FileNotFoundError: [Errno 2] No such file or directory:
Issue -
State: closed - Opened by RunCode90 over 6 years ago
#55 - The questions about training
Issue -
State: open - Opened by song-heng over 6 years ago
- 11 comments
#54 - Why do you use sum(decode_lengths) to as the count to update losses and topkaccuracies?
Issue -
State: open - Opened by dwang68 over 6 years ago
- 3 comments
#53 - Low volatile GPU-Util but high GPU Memory Usage
Issue -
State: closed - Opened by yolo615 over 6 years ago
- 13 comments
#52 - update readme with instructions for BLEU-1 to BLEU-4
Pull Request -
State: closed - Opened by kmario23 over 6 years ago
- 1 comment
#51 - RuntimeError Expected object of type torch.cuda.FloatTensor but found type torch.FloatTensor for argument #4 'other'
Issue -
State: closed - Opened by JohnComeon over 6 years ago
- 6 comments
#50 - The question about the vocabulary size.
Issue -
State: open - Opened by zkself over 6 years ago
- 7 comments
#49 - How to recover training at a check point
Issue -
State: closed - Opened by giangnguyen2412 over 6 years ago
- 1 comment
#48 - RuntimeError:Unable to create link(name already exists)
Issue -
State: closed - Opened by JohnComeon over 6 years ago
- 3 comments
#47 - Pre-trained word embeddings
Issue -
State: closed - Opened by DWhettam over 6 years ago
- 1 comment
#46 - SourceChangeWarning
Issue -
State: closed - Opened by 672401341 over 6 years ago
- 9 comments
#45 - Converting to ONNX?
Issue -
State: closed - Opened by ghost over 6 years ago
#44 - usage
Issue -
State: closed - Opened by JohnComeon over 6 years ago
- 2 comments
#43 - lstmcell GPU OOM
Issue -
State: closed - Opened by ghost over 6 years ago
- 1 comment
#42 - The steps to run the model from preprocessing to inference are really vague
Issue -
State: closed - Opened by giangnguyen2412 over 6 years ago
- 1 comment
#41 - regarding attention concatentation in beam search
Issue -
State: closed - Opened by fawazsammani over 6 years ago
- 1 comment
#40 - why are you using pack padded seq if you've already removed the pads in inputs?
Issue -
State: closed - Opened by homelifes over 6 years ago
- 1 comment
#39 - Pretrained model?
Issue -
State: closed - Opened by PabloRR100 over 6 years ago
- 2 comments
#38 - Why do you use log_softmax in sampling?
Issue -
State: open - Opened by RodinIvan over 6 years ago
- 1 comment
#37 - How to do calculate all bleu scores during evaluvaion
Issue -
State: closed - Opened by HarryDresden1 over 6 years ago
- 9 comments
#36 - regarding the batch_size_t
Issue -
State: closed - Opened by fawazsammani over 6 years ago
- 2 comments
#35 - are the testing images used?
Issue -
State: closed - Opened by fawazsammani over 6 years ago
- 7 comments
#34 - hello,i can't find the pre-trained model you provided "BEST_checkpoint_coco_5_cap_per_img_5_min_word_freq.pth.tar"
Issue -
State: closed - Opened by RogerNEU over 6 years ago
- 3 comments
#33 - Sampling method not following the paper?
Issue -
State: closed - Opened by PabloRR100 over 6 years ago
- 3 comments
#32 - Some trouble in the model.py about fine-tuning
Issue -
State: closed - Opened by czhxiaohuihui over 6 years ago
- 1 comment
#31 - fix corpus_bleu error
Pull Request -
State: closed - Opened by anotherTK over 6 years ago
- 1 comment
#30 - Predicting to a new image (Teacher Forcing)
Issue -
State: closed - Opened by ricoms over 6 years ago
- 2 comments
#29 - fix bleu evaluation for latest nltk
Pull Request -
State: closed - Opened by jason718 over 6 years ago
- 1 comment
#28 - Regarding the Attention Addition
Issue -
State: closed - Opened by fawazsammani over 6 years ago
- 2 comments
#27 - Attension Network explanation?
Issue -
State: open - Opened by roachsinai over 6 years ago
- 11 comments
#26 - Question about reference
Issue -
State: closed - Opened by RenjieWei over 6 years ago
- 1 comment
#25 - some trouble with the input files.
Issue -
State: closed - Opened by Goody7 over 6 years ago
- 3 comments
#24 - How to generate sentid's for another dataset?
Issue -
State: open - Opened by hughkf over 6 years ago
- 1 comment
#23 - ValueError: 'arr' does not have a suitable array shape for any mode
Issue -
State: closed - Opened by holibert over 6 years ago
- 1 comment
#22 - using original caption dataset for flickr8k & flickr30k
Issue -
State: open - Opened by ajengwulandari over 6 years ago
#21 - Last layer activation
Issue -
State: closed - Opened by miroblog over 6 years ago
- 2 comments
#20 - predict linear layer's input is just hidden states but in original paper, they combined with [L(word_embed+Wh+Uc)]
Issue -
State: open - Opened by joelxiangnanchen over 6 years ago
- 3 comments
#19 - corpus_bleu() got an unexpected keyword argument 'emulate_multibleu'
Issue -
State: closed - Opened by norahsakal over 6 years ago
- 4 comments
#18 - error in computing the beam search for all images
Issue -
State: closed - Opened by fawazsammani over 6 years ago
- 11 comments
#17 - regarding the bleu score reported
Issue -
State: closed - Opened by fawazsammani almost 7 years ago
- 4 comments
#16 - About the field image caption
Issue -
State: closed - Opened by ZHUXUHAN almost 7 years ago
- 4 comments
#15 - TypeError: can't pickle _thread._local objects
Issue -
State: closed - Opened by cpaaax almost 7 years ago
- 5 comments
#14 - Loss function in Stochastic "Hard" Attention
Issue -
State: open - Opened by kouui almost 7 years ago
- 1 comment
#13 - what is the lowest loss we can achieve with this architecture?
Issue -
State: open - Opened by laptopmutia almost 7 years ago
- 3 comments
#12 - What is the function of sample captions and captions per image
Issue -
State: closed - Opened by laptopmutia almost 7 years ago
- 6 comments
#11 - A question beyond the code itself
Issue -
State: closed - Opened by RobinWenqian almost 7 years ago
- 10 comments
#10 - are all the functions used?
Issue -
State: closed - Opened by fawazsammani almost 7 years ago
- 7 comments
#9 - pretrained models damaged
Issue -
State: closed - Opened by fawazsammani almost 7 years ago
- 4 comments
#8 - How to assign certain gpu(s)to train the model?
Issue -
State: closed - Opened by RobinWenqian almost 7 years ago
- 8 comments
#7 - Some problem in creat_input_files.py
Issue -
State: closed - Opened by RobinWenqian almost 7 years ago
- 6 comments
#6 - why to use teacher-forcing in function Validate()?
Issue -
State: closed - Opened by wyk-nku almost 7 years ago
- 4 comments
#5 - Do we need to use word2vec?
Issue -
State: closed - Opened by chaoyan1037 almost 7 years ago
- 3 comments