ruotianluo/ImageCaptioning.pytorch issues and pull requests

#91 - Does this code use word2vec to generate the word embeddings?

Issue - State: open - Opened by KahHwa over 5 years ago

#90 - Where is your resnet 101 Link?

Issue - State: closed - Opened by he159ok over 5 years ago - 4 comments

#89 - Bug fix for exception during training when `tf` is not `None` and `opt.language_eval` is `0`

Pull Request - State: closed - Opened by gtbai over 5 years ago

#88 - Show and tell model and diverse beam search.

Issue - State: closed - Opened by mememimis over 5 years ago

#87 - RuntimeError: Expected object of type torch.cuda.FloatTensor but found type torch.FloatTensor for argument #4 'other'

Issue - State: open - Opened by HAWLYQ over 5 years ago - 1 comment

#86 - Fix bug causing beams to not be reordered by log probability.

Pull Request - State: open - Opened by daphnei over 5 years ago

#85 - where in the code did you extract the fake_region, conv_feat, conv_feat_embed from the image?

Issue - State: open - Opened by homelifes almost 6 years ago - 4 comments

#84 - Possible Memory leak for RAM MEM

Issue - State: open - Opened by cookielee77 almost 6 years ago - 4 comments

#83 - Model predicts 10 captions even when there is only 1 image in the folder

Issue - State: open - Opened by KahHwa almost 6 years ago - 3 comments

#82 - LSTMcell problem

Issue - State: open - Opened by Daic115 almost 6 years ago

#81 - training speed

Issue - State: open - Opened by Iamal1 almost 6 years ago - 4 comments

#80 - my cocotalk_att/feats_att.h5 file is 369 GB,while the readme said it's about 200GB. Is anything wrong here

Issue - State: closed - Opened by li10141110 about 6 years ago - 1 comment

#79 - what loss function you used for training those models

Issue - State: closed - Opened by SkylerZheng about 6 years ago - 2 comments

#78 - training topdown model log_st/model file or directory not found

Issue - State: closed - Opened by KahHwa about 6 years ago

#77 - Error while running eval.py on pretrained models

Issue - State: open - Opened by KahHwa about 6 years ago - 4 comments

#76 - Faster-RCNN features not used when training Top-down Bottom up model?

Issue - State: open - Opened by ramprs about 6 years ago

#75 - Grad for student forcing in training

Issue - State: closed - Opened by qy-feng about 6 years ago - 1 comment

#74 - Error while fine-tuning

Issue - State: closed - Opened by metro-smiles about 6 years ago - 4 comments

#73 - Error while training on a new dataset

Issue - State: closed - Opened by metro-smiles about 6 years ago - 1 comment

#72 - Fix eval script with python3 and pytorch0.4

Pull Request - State: closed - Opened by dmitriy-serdyuk about 6 years ago

#71 - Pytorch 4

Pull Request - State: closed - Opened by dmitriy-serdyuk about 6 years ago

#70 - Bump pytorch to 4.1

Pull Request - State: closed - Opened by dmitriy-serdyuk about 6 years ago - 1 comment

#69 - Upgrade code to PyTorch 0.4

Issue - State: closed - Opened by metro-smiles about 6 years ago - 9 comments

#68 - How to feed fc_feats at 1st time step in show attend and tell model

Issue - State: closed - Opened by jamiechoi1995 about 6 years ago - 1 comment

#67 - What is att_feats?

Issue - State: closed - Opened by LaiPiXiong about 6 years ago - 3 comments

#66 - How to train attmodel?

Issue - State: closed - Opened by LaiPiXiong about 6 years ago - 9 comments

#65 - what's p_att_feats mean?

Issue - State: closed - Opened by LaiPiXiong about 6 years ago

#64 - If I don't use pre-trained att-features, how can I get att-features?

Issue - State: closed - Opened by LaiPiXiong about 6 years ago

#63 - Extract FC feature in resnet_utlis.py

Issue - State: open - Opened by jamiechoi1995 about 6 years ago - 2 comments

#62 - ImportError: No module named misc.resnet_utils

Issue - State: closed - Opened by jamiechoi1995 over 6 years ago - 1 comment

#61 - FileNotFoundError: No such file or directory: 'log_st/model.pth'

Issue - State: closed - Opened by erkams over 6 years ago

#60 - I met an error during training

Issue - State: closed - Opened by fearless77 over 6 years ago - 1 comment

#59 - How to achieve the high performance?

Issue - State: closed - Opened by ivy94419 over 6 years ago - 10 comments

#58 - I got an error during training process

Issue - State: closed - Opened by fearless77 over 6 years ago - 1 comment

#57 - failed to generate and save fc and att features to .h5 files, for my own datasets

Issue - State: closed - Opened by SkylerZheng over 6 years ago - 3 comments

#56 - the paper pointed that fc is generated by faster rcnn, I find you use resnet instead.

Issue - State: closed - Opened by BetterZhouXu over 6 years ago - 4 comments

#55 - Error in eval

Issue - State: closed - Opened by dhimanpd over 6 years ago - 4 comments

#54 - where is infos_pkl

Issue - State: closed - Opened by wanglinteng over 6 years ago - 1 comment

#53 - I cannot reach the score where "readme.md“ mentioned.

Issue - State: closed - Opened by LaiPiXiong over 6 years ago - 5 comments

#52 - When beam_size > 1, it shows size mismatch error.

Issue - State: closed - Opened by LaiPiXiong over 6 years ago

#51 - About the batch_size

Issue - State: closed - Opened by congve1 over 6 years ago - 1 comment

#50 - It generate all the same sentences to different images, and i found the input of LSTM in the second time step is all zeros.

Issue - State: closed - Opened by LaiPiXiong over 6 years ago - 7 comments

#49 - ValueError: sampler should be an instance of torch.utils.data.Sampler.

Issue - State: closed - Opened by zhjohnchan over 6 years ago - 2 comments

#48 - Generate soft attention pictures of each word

Issue - State: open - Opened by Silence1995 over 6 years ago - 1 comment

#47 - Train on Flicker8k Dataset

Issue - State: open - Opened by ireneMsm2020 over 6 years ago - 2 comments

#46 - What are the abbreviation of "fc" and "att" meaning?

Issue - State: closed - Opened by thumbe12856 over 6 years ago - 6 comments

#45 - It slowed down as training processed

Issue - State: closed - Opened by LaiPiXiong over 6 years ago - 1 comment

#44 - Performance of CIDEr decreases when performing self critic training at first 6000 iterations

Issue - State: closed - Opened by vanpersie32 over 6 years ago

#43 - if i want to train cnn and LSTM，how should i do????

Issue - State: closed - Opened by 123moon over 6 years ago - 5 comments

#42 - Cpu issue

Issue - State: open - Opened by soheylsa over 6 years ago - 10 comments

#41 - Cpu issue

Issue - State: closed - Opened by soheylsa over 6 years ago

#40 - about topdown model

Issue - State: open - Opened by jiasureZ over 6 years ago - 3 comments

#39 - Teacher Forcing

Issue - State: open - Opened by karanchahal almost 7 years ago

#38 - saving to and loading from single h5

Pull Request - State: closed - Opened by ReyhaneAskari almost 7 years ago - 12 comments

#37 - Pre trained vectors

Issue - State: open - Opened by karanchahal almost 7 years ago - 2 comments

#36 - Python-3 support

Issue - State: open - Opened by kmario23 almost 7 years ago - 9 comments

#35 - UnboundLocalError: local variable 'cnn_optimizer' referenced before assignment

Issue - State: closed - Opened by JaneLou almost 7 years ago

#34 - Performance

Issue - State: closed - Opened by Liu0329 almost 7 years ago - 1 comment

#33 - I met a error in the training...

Issue - State: open - Opened by weiweili123 almost 7 years ago - 8 comments

#32 - We can help each other! Finding friend who is studying image captioning , my WeChat ID lijingx-正在研究image caption 朋友们来一起讨论代码

Issue - State: open - Opened by li10141110 almost 7 years ago

#31 - About att_feature

Issue - State: open - Opened by ssdzxq almost 7 years ago - 2 comments

#30 - About using cpu on evaluating

Issue - State: open - Opened by xxbbml almost 7 years ago

#29 - The attention mechanism of '''top-down'' is different from that of ''show attend and tell'', but they seem the same in your code.

Issue - State: closed - Opened by ghost about 7 years ago - 4 comments

#28 - Remove `if True:` in train.py

Pull Request - State: open - Opened by lsqshr about 7 years ago

#27 - A question about evaluation funciton

Issue - State: closed - Opened by lxtGH about 7 years ago - 2 comments

#26 - performances of each model and self-critical

Issue - State: closed - Opened by gctian about 7 years ago - 4 comments

#25 - questions about initializing the lstm hidden states

Issue - State: closed - Opened by brisker about 7 years ago - 16 comments

#24 - Update train.py

Pull Request - State: open - Opened by JyiHUO about 7 years ago

#23 - Write to one npy

Pull Request - State: closed - Opened by dmitriy-serdyuk about 7 years ago - 7 comments

#22 - Features saved in many npy files are slow to read

Issue - State: open - Opened by dmitriy-serdyuk about 7 years ago - 11 comments

#21 - how to control the num of thread?

Issue - State: closed - Opened by qingchenwuhou about 7 years ago - 2 comments

#20 - One problem during the training process

Issue - State: closed - Opened by lxtGH about 7 years ago - 8 comments

#19 - Some confusion about adaptive attention model

Issue - State: closed - Opened by a380922457 about 7 years ago - 4 comments

#18 - Multi-GPU Training Support

Issue - State: open - Opened by ahkarami about 7 years ago - 5 comments

#17 - TensorBoard Problem

Issue - State: closed - Opened by ahkarami about 7 years ago - 2 comments

#16 - A bug in 'fc' model when using GRU?

Issue - State: closed - Opened by danieljf24 about 7 years ago - 2 comments

#15 - Train on other dataset

Issue - State: open - Opened by ahkarami about 7 years ago - 18 comments

#14 - Is it possible to detect anomaly with neuraltalk?

Issue - State: open - Opened by bemoregt about 7 years ago

#13 - Potential Issue of using multi-gpu

Issue - State: closed - Opened by chihyaoma about 7 years ago - 2 comments

#12 - AttributeError: 'Namespace' object has no attribute 'use_att'

Issue - State: closed - Opened by demobin8 about 7 years ago - 8 comments

#11 - UnboundLocalError: local variable 'resnet' referenced before assignment

Issue - State: closed - Opened by demobin8 about 7 years ago - 1 comment

#10 - Benchmarks

Issue - State: open - Opened by ruotianluo over 7 years ago - 71 comments

#9 - Evaluation: AttributeError: 'Namespace' object has no attribute 'caption_model'

Issue - State: closed - Opened by danjcosg over 7 years ago - 14 comments

#8 - fix bug

Pull Request - State: closed - Opened by raoyongming over 7 years ago - 1 comment

#7 - dataloaderraw.py, line 25 - IOError: [Errno 2] No such file or directory: '/home-nfs/rluo/rluo/model/pytorch-resnet/resnet101.pth'

Issue - State: closed - Opened by danjcosg over 7 years ago - 1 comment

#6 - Update resnet feature extraction: torch.mean()

Pull Request - State: closed - Opened by gujiuxiang over 7 years ago

#5 - Doubt in criterion

Issue - State: closed - Opened by ayushgithub over 7 years ago - 1 comment

#4 - some bug found in using

Issue - State: closed - Opened by hitluobin over 7 years ago - 22 comments

#3 - some code missing?

Issue - State: closed - Opened by cswhjiang over 7 years ago - 37 comments

#2 - Update readme

Pull Request - State: closed - Opened by gujiuxiang over 7 years ago - 2 comments

#1 - The performance on MS COCO Val5000

Issue - State: closed - Opened by eriche2016 almost 8 years ago - 6 comments

GitHub / ruotianluo/ImageCaptioning.pytorch issues and pull requests