issues
search
airsplay
/
lxmert
PyTorch code for EMNLP 2019 paper "LXMERT: Learning Cross-Modality Encoder Representations from Transformers".
MIT License
923
stars
157
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
VQA: Split-wise result during each epoch
#71
gkv91
opened
4 years ago
0
number of data characters(457269) cannot be 1 more than a multiple of 4
#70
YangYL18
opened
4 years ago
3
Issue about the weights of cross attention in x_layers
#69
XChuanLee
opened
4 years ago
2
Asmita lxmert activation
#68
AsmitaBhat30
closed
4 years ago
0
Question about GQA accuracy.
#67
simpleshinobu
closed
4 years ago
1
how to know the object name from obj id in vg_gqa_imgfeat?
#66
luomancs
closed
4 years ago
2
Question about VG Caption
#65
VickiCui
closed
4 years ago
1
Non-docker way to extract image features?
#64
johntiger1
opened
4 years ago
2
More than 5 captions per image for COCO data
#63
j-min
closed
4 years ago
1
'\n' characters in sentences in *.json data files.
#62
j-min
opened
4 years ago
3
Shapes of inputs and output in entry.py
#61
ecekt
closed
4 years ago
2
entry.py model loading
#60
ecekt
closed
4 years ago
2
VISUAL_CONFIG for Epoch20_LXRT.pth
#59
ecekt
opened
4 years ago
2
The Hyper parameters for the VizWiz datasets
#58
runzeer
opened
4 years ago
10
How to generate the multi-model fusion results
#57
runzeer
opened
4 years ago
7
QA data used in image-QA ablation study
#56
j-min
closed
4 years ago
2
The heatmap shown in the slides
#55
doppelbeaver
opened
4 years ago
1
Dev-set splits in paper
#54
j-min
closed
4 years ago
2
Making this available at hugging face models directory?
#53
vibhavagarwal5
closed
4 years ago
3
GQA submission
#52
zaynmi
opened
4 years ago
18
masked_lm_loss being optimised even when label is not matched ?
#51
singh-hrituraj
closed
4 years ago
2
Could you share the finetuning time cost and the device information?
#50
moyutianque
closed
4 years ago
2
VQA finetuning training time
#49
johntiger1
closed
4 years ago
6
Can not use multiply gpus for bottom-up feature extraction
#48
Wangt-CN
opened
4 years ago
6
"pre-training" section in the readme
#47
johntiger1
opened
4 years ago
2
seed
#46
runzeer
closed
4 years ago
0
issues when finetuning the pretrained models for GQA
#45
runzeer
opened
4 years ago
1
Broken link for Faster-RCNN model for Feature Extraction
#44
shubhamagarwal92
closed
4 years ago
2
How to build 'all_ans.json'
#43
sileod
closed
4 years ago
1
cudaSuccess (9 vs. 0)
#42
volquelme
closed
4 years ago
1
Why does object and attribute loss not be masked?
#41
forjiuzhou
opened
4 years ago
9
Could you also provide the fine-tuned model?
#40
Princesston
closed
4 years ago
1
Pretraining attribute and object loss don't match the ones in the log file
#39
kritiagg
opened
4 years ago
12
Empty file in pretraining VQA
#38
KfirBernstein
opened
4 years ago
1
CPU memory usage is too high and other queries
#37
prajjwal1
closed
4 years ago
19
Pre-training doesn't work
#36
claudiogreco
closed
4 years ago
5
Would you like to share all the checkpoints of the model during the training phase?
#35
JizeCao
opened
4 years ago
1
When i use your docker image to extract butd features, should i need to make pycaffe or build Caffe?
#34
Zhao-Yuting
closed
4 years ago
2
FP16 for training acceleration
#33
zhao1iang
closed
4 years ago
2
problem in feature extraction
#32
luoruixuan
closed
4 years ago
2
RuntimeError: cuda runtime error (38) : no CUDA-capable device is detected at /opt/conda/conda-bld/pytorch_1549630534704/work/aten/src/THC/THCGeneral.cpp:51
#31
Zzmonica
closed
4 years ago
2
Low training speed
#30
ypengc7512
closed
4 years ago
1
AttributeError: 'NoneType' object has no attribute 'init_bert_weights'
#29
AIstudentSH
opened
4 years ago
14
CUDA error: invalid configuration argument on large
#28
twistedcubic
opened
4 years ago
2
How to extract the CrossModality Output
#27
256785
closed
4 years ago
1
What are the exact 4-dimension spatial features for ROIs?
#26
JizeCao
closed
4 years ago
1
Visual genome features
#25
JizeCao
closed
4 years ago
2
oov of gqa answer (gqa/trainval_ans2label.json only has 1842 answers)
#24
LinkToPast1990
opened
4 years ago
4
How can I reproduce your pre-training result in EMNLP paper?
#23
jkhong9
closed
4 years ago
2
Can you provide a log file of lxmert pretraining
#22
ParadoxZW
opened
4 years ago
8
Previous
Next