microsoft / SwinBERT

Research code for CVPR 2022 paper "SwinBERT: End-to-End Transformers with Sparse Attention for Video Captioning"

https://arxiv.org/abs/2111.13196

microsoft/SwinBERT Issues

Missing swin_small.py
Closed 4 months ago1
149.4 vs 120.6 for CIDEr on MSVD dataset
Closed 6 months ago2
Trimming the YouCook2 videos + "./datasets/YouCook2/yc2_subtitles.jsonl"?
Updated 6 months ago4
Missing Caption Files For YouCook2 Dataset
Updated 6 months ago1
No such file or directory: 'datasets/VATEX/test.img.tsv'
Closed 6 months ago
docker image link can not be used.
Closed 6 months ago
Combining Features of Swin Transformer and other Features
Updated 7 months ago
attention_scores= attention_scores + attention_mask
Updated 8 months ago1
raise KeyError(key) from None KeyError: 'RANK'
Updated 8 months ago1
程序运行
Updated 8 months ago1
FileNotFoundError: [Errno 2] No such file or directory: 'datasets/MSVD/frame_tsv/train_32frames.img.tsv'
Updated 8 months ago4
anaconda environment instead of Docker
Updated 9 months ago2
Issue arise training on HPC
Updated 9 months ago
Using Vatex dataset to train my own chinese data
Updated 9 months ago2
Is there any other settings? I run evaluation code directly always gets very low scores
Closed 10 months ago
Training problems when reproduce the results of MSRVTT dataset
Updated 10 months ago17
Can you share the multi-GPU training command?
Closed 2 years ago5
evalcap preparation
Closed 2 years ago2
How to download the raw videos of YouCook2?
Updated a year ago1
Is there any parameter to tweak in the code to get longer captions?
Updated a year ago
Inference error with CPU
Updated a year ago2
gpu memory requirements?
Closed 2 years ago1
Do the input lengths have to be fixed during training?
Updated a year ago
What are the entire packages and dependencies for running the code? Is someone run the code in local conda environment successfully?
Closed a year ago1
When will you release the tutorial for frame-based TSV generation? Thanks!
Updated a year ago11
Train MSVD dataset using VATEX pretrained model? Thanks
Updated a year ago1
oserror (errno 30) read-only file system python docker
Closed 2 years ago2
Could you please provide the file msvd-retrieval_train-val-test.json
Closed 2 years ago1
How can I get frame_tsv/train_32frames.img.lineidx for MSVD?
Updated a year ago3
Questions about the value of 'loss_sparse_w' in command
Updated a year ago1
Questions About Frames extracting
Updated a year ago2
Inconsistent reproduced results with the results reported in the paper
Closed 2 years ago17
Data split re MSVD and MSRVTT caption
Updated a year ago1
Questions about masked_loss_img
Updated 2 years ago
Can you directly provide the running environment of the code？ thank you！
Updated 2 years ago1
container_abcs
Updated 2 years ago
The performance of adopting the Video Swin Transformer as offline-extracted video encoder?
Updated 2 years ago2
About Attention Mask
Updated 2 years ago1
Question about learning rate in Multi-gpu training
Updated 2 years ago1
Adapt to custom datasets
Updated 2 years ago
Training my own dataset, but the indices equal 0 and in the output path (coco json format) don't have any caption.
Updated 2 years ago
请问该代码支持多块gpu训练吗？
Closed 2 years ago3
Using my own data to train
Updated 2 years ago
One data typo
Updated 2 years ago1
TVC Dataset
Updated 2 years ago1
The link to raw videos
Updated 2 years ago2
can't download pretrained models from the https://datarelease.blob.core.windows.net/swinbert/
Updated 2 years ago4
A file is not found: no module named 'src.evalcap.coco_caption'
Closed 2 years ago3
Can you share the train-val-test.json
Closed 2 years ago2
couldn't find msrvtt-32frm.zip
Closed 2 years ago1