issues
search
happyharrycn
/
actionformer_release
Code release for ActionFormer (ECCV 2022)
MIT License
415
stars
77
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
multigpu training
#88
AndongDeng
closed
1 year ago
2
Error "predicted offsets must be non-negative"
#87
thanhhff
closed
1 year ago
1
Assertion error
#86
Subhraneil
closed
1 year ago
0
不连续时间帧的回归问题
#85
tingfengwanxiao
closed
1 year ago
3
Understanding gt_cls_labels
#84
tvaranka
closed
1 year ago
4
local window size
#83
shiyi-z
closed
1 year ago
2
Negative Action Regions (in Epic Kitchens)
#82
beasteers
closed
1 year ago
2
About EgoVLP feature for Ego4D.
#81
EdenGabriel
closed
1 year ago
7
Hello, the author, I don't know whether your experiment has the result of recall rate or AR@AN
#80
jiaxibei323
closed
1 year ago
3
compute cls loss
#79
OpenSorceYCW
closed
1 year ago
13
Training without gt_segments, gt_labels.
#78
grayfactory
closed
1 year ago
2
error
#77
OpenSorceYCW
closed
1 year ago
2
About thumos14 features using TSP
#76
weike382
closed
1 year ago
6
Measure GMACs/FLOPS
#75
TuanTNG
closed
1 year ago
2
nms_1d_cpu
#74
nshah0812
closed
1 year ago
2
feat_offset
#73
jzhao004
closed
1 year ago
2
import nms_1d_cpu ImportError
#72
JiangZhouHao
closed
1 year ago
4
Question about regression ranges
#71
zivnachum
closed
1 year ago
5
About extracting feature.
#70
EdenGabriel
closed
1 year ago
2
Question about the choice of regression_range.
#69
QX-N
closed
1 year ago
3
About n_mha_win_size in eval.
#68
EdenGabriel
closed
1 year ago
4
why set num_classes as 1 in ActivityNet dataset module?
#67
fake-warrior8
closed
1 year ago
2
Visualization code
#66
minghu0830
closed
1 year ago
3
What is the test video number of Thumos14 dataset?
#65
fake-warrior8
closed
1 year ago
2
what is the h5py version?
#64
fake-warrior8
closed
1 year ago
0
about win_size
#63
nqxjzm
closed
1 year ago
2
Why are the reported results of TadTR on Thumos14 and ActivityNet quite different with those in TadTR paper?
#62
ttgeng233
closed
1 year ago
2
How to improve？
#61
QingLuogwj
closed
1 year ago
1
Question about such strong I3D feature
#60
yangmin666
closed
1 year ago
6
How to reproduce results using i3d for ActivityNet?
#59
g1910
closed
1 year ago
8
Plan to release Ego4d trained model
#58
HYUNJS
closed
1 year ago
2
ActivityNet trainning json file
#57
houzhijian
closed
1 year ago
4
Training on Own Dataset - on short videos
#56
rafiko1
closed
1 year ago
6
Merge rgb and flow?
#55
bleakie
closed
1 year ago
8
How to modify regression range if I change the pyramid levels?
#54
ttgeng233
closed
1 year ago
4
multi GPU training
#53
shiyi-z
closed
1 year ago
2
Figure 3. Visualization of our results.
#52
jiaxibei323
closed
1 year ago
3
How to get the action labels?
#51
shakib-root
closed
1 year ago
4
Does it only support the gt_segment that just has one action for one moment?
#50
ttgeng233
closed
1 year ago
6
With BiFPN
#49
zhangzhili1112
closed
2 years ago
4
Why disabling score fusion get even better results?
#48
makecent
closed
2 years ago
4
add the first verson custom dataset.md
#47
tzzcl
opened
2 years ago
2
Computing mPA
#46
fsardari
closed
2 years ago
2
Hello, I can't open the google drive link you shared about ActivityNet 13. Pre-trained Model I don't know if there is a Baidu cloud link or other ways to open the link?
#45
jiaxibei323
closed
2 years ago
1
How to visualize the results like in the paper
#44
jiaxibei323
closed
2 years ago
2
Can this method be used for multi-label temporal localization?
#43
ttgeng233
closed
2 years ago
4
in label_points_single_video, why do u select min_len index to build regression target?
#42
freeman-1995
closed
2 years ago
2
in anet.py dataset, why u need to minus 0.5*num_frames? this is weird
#41
freeman-1995
closed
2 years ago
2
ActivityNet 1-class vs. 200-class problem
#40
egoodman92
closed
2 years ago
2
PointGenerato
#39
QingLuogwj
closed
2 years ago
6
Previous
Next