batra-mlp-lab/visdial-challenge-starter-pytorch
Starter code in PyTorch for the Visual Dialog challenge
PythonBSD-3-Clause
Issues
- 0
why do max_sequence_length - 1 in dataset.py
#53 opened by Thinking-more - 0
- 0
How to finetune the model (generated by train.py) with dense annotations (train + val)?
#40 opened by dongliangchang - 0
- 0
generative decoder
#35 opened by weichentangming - 1
Bounding box coordinates
#24 opened by shubhamagarwal92 - 6
Training step is too slow
#25 opened by guoyang9 - 0
Torch=1.0.0 is not found
#33 opened by abrar1515 - 3
About concat_history in dataset.py
#32 opened by gicheonkang - 5
- 2
Late Fusion demo not working
#29 opened by shubhamagarwal92 - 0
- 1
- 0
- 2
- 13
About image features
#15 opened by JXZe - 4
Need suggestion about embeddings
#17 opened by shubhamagarwal92 - 8
Tokenizing is slow
#13 opened by idansc - 3
generative decoder
#4 opened by angelboy000 - 1
Extracting Actual Images
#14 opened by SanatSharma - 2
extract image features
#12 opened by wh0330 - 1
visdial_1.0_val_dense_annotations.json
#11 opened by angelboy000 - 2
- 3
- 4
How to get multi-gpu to work?
#8 opened by hsm207 - 7
ffi.lua:56 expected align(#) on line 579
#2 opened by bedsDev - 2
VisDial v0.9
#3 opened by choys73 - 1
Softmax dimension is wrong.
#1 opened by hwanheelee1993