issues
search
taoyang1122
/
adapt-image-models
[ICLR'23] AIM: Adapting Image Models for Efficient Video Action Recognition
Apache License 2.0
278
stars
21
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
Test results
#48
Liu-arch
opened
5 months ago
0
model下载问题
#47
Douyi2333
opened
8 months ago
0
I would like to know what testing protocol the 88.9% on Diving48 is based on.
#46
Changwei-Ouyang
opened
12 months ago
0
An error with training on multiple GPUs
#45
wlsrick
opened
1 year ago
1
A Error in the Running the Train Code.
#44
YeungChiu
opened
1 year ago
0
The issue of being unable to continue training with already trained weights.
#43
Xfold-masker
opened
1 year ago
0
Have the authors tried fine-tuning the parameters of the LayerNorm layer with it turned on? If so, what were the results?
#42
Changwei-Ouyang
closed
12 months ago
1
visualizing code
#41
Changwei-Ouyang
closed
12 months ago
0
How to finetune?
#40
smandava98
opened
1 year ago
1
The training set and val set in the Diving48 dataset
#39
Changwei-Ouyang
closed
1 year ago
1
The dimension of vit_clip and vit_imagenet
#38
wlsrick
opened
1 year ago
2
About learning rate setting
#37
hsi-che-lin
closed
1 year ago
1
Weights Initialization
#36
liyi-ff
opened
1 year ago
1
About the performance gap with the released checkpoints
#35
yangbang18
opened
1 year ago
6
Train the ViT in 4 2080Ti
#34
wlsrick
opened
1 year ago
1
Encoding error while loading val data on k400
#33
TJQdoIt9527
closed
1 year ago
1
About hyperparameters setting for reproducing the results
#32
hsi-che-lin
closed
1 year ago
3
Inquiring about the number of views during test time.
#31
Backdrop9019
opened
1 year ago
1
About GPU memory usage
#30
Zhc4zzz
opened
1 year ago
1
It seems that the training memory is not reduced
#29
casillas1111
opened
1 year ago
1
About CLIP-based adapter
#28
kaiw7
opened
1 year ago
1
Training effiency
#27
Fanye12
closed
1 year ago
2
Question about the Custom loss function on Recognition (like ViT model)
#26
andrewwang0612
opened
1 year ago
1
Can't reproduce Something-Somethingv2 training and results
#25
or7r
opened
1 year ago
7
something questions about hyper parameters and experiment results
#24
007invictus
opened
1 year ago
1
About training time.
#23
leexinhao
closed
1 year ago
2
mmaction
#22
dmmSJTU
opened
1 year ago
1
why the second dimension is n*b?
#21
007invictus
closed
1 year ago
4
Did you modify any config when change the frame_len from 8 to 16 or 32?
#20
leexinhao
closed
1 year ago
1
pretrained model
#19
dmmSJTU
opened
1 year ago
7
swin2d_base.py and swin2d_adapter_base.py
#18
XiudingCai
opened
1 year ago
1
Where is the function “init weights” called
#17
xuesongnie
closed
1 year ago
1
T-MSA and S-MSA
#16
chaolistat
closed
1 year ago
3
SSv2 models
#15
or7r
closed
1 year ago
2
How can we freeze the layer?
#14
VoyageWang
closed
1 year ago
2
Could you please share the swin transformer model weights on Kinetics 400?
#13
K0145vin
closed
1 year ago
2
where you load the checkpoint of ViT backbone?
#12
SHIBOYA
closed
1 year ago
2
Maybe you should set `find_unused_parameters=True` in you config file?
#11
leexinhao
closed
1 year ago
4
where is your model?
#10
SHIBOYA
closed
1 year ago
1
Visualization Experiments
#9
laohuijiadezhu
closed
1 year ago
1
在ViT_CLIP 中没有发现冻结层不做训练的代码，请问您这边在哪里做了处理，就是冻结预训练好的bcakbond权重
#8
xiaoyaod
closed
1 year ago
2
Training log available?
#7
simonJJJ
closed
1 year ago
5
About input frames and sampling interval
#6
BinhuiXie
closed
1 year ago
1
Can't reproduce Diving48 training
#5
ahnGeo
closed
1 year ago
1
Do you have pretrained model weights like ViT-S/16 or ViT-S/16？
#4
RayTang88
closed
1 year ago
1
Do you have pretrained model weights like ViT-S/16 or ViT-S/16？
#3
RayTang88
closed
1 year ago
0
Enquiry on the accessibility of SSV2 configs
#2
PeiqinZhuang
closed
1 year ago
1
requirements.txt natplotlib => matplotlib?
#1
HolyDragon57
closed
1 year ago
1