issues
search
microsoft
/
SimMIM
This is an official implementation for "SimMIM: A Simple Framework for Masked Image Modeling".
https://arxiv.org/abs/2111.09886
MIT License
917
stars
86
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
About mask_token in SimMIM codes
#41
YangGangZhiQi
opened
8 months ago
0
inverse swin
#40
betterze
opened
12 months ago
1
用自监督结果的预训练权重对下游任务精度影响
#39
swjtulinxi
opened
1 year ago
0
How can I solve this problem
#38
haibo12
opened
1 year ago
2
Could you provide the pretrained ViT-Base model with patch size 16？
#37
irsLu
opened
1 year ago
0
Setting for Linear eval
#36
tonysy
opened
1 year ago
0
Tt
#35
Kingris9090
closed
1 year ago
0
any plan to release the code of DDP training of swinV2-G with multi machine？
#34
dongzhiwu
opened
2 years ago
0
Questions about AvgDist
#33
yjsunnn
opened
2 years ago
0
Any plan to support 3D version?
#32
james20141606
opened
2 years ago
0
The specific configs and code for downstream tasks, like semantic segmentation and object detection
#31
Haoqing-Wang
opened
2 years ago
0
Why Swin-Large-W12 contains [36, 36] `encoder.layers.3.blocks.0.attn.relative_position_index`
#30
nijkah
opened
2 years ago
1
Indentation bug in utils.remap_pretrained_keys_vit
#29
BradMcDanel
opened
2 years ago
0
Performance using the cosine distance
#28
LiyaoTang
opened
2 years ago
0
Train from scratch
#27
naoki7090624
opened
2 years ago
1
Why there is "no_weight_decay" function for Swin-T but not for VIT
#26
TongZhangTHU
opened
2 years ago
0
Could you please release the pre-trained R50 model?
#25
liming-ai
opened
2 years ago
0
For cnn architecture like resnet50
#24
Rui-Zhou-2
opened
2 years ago
0
a bug in loading pretrain checkpoint ， in utils.py line 118， you dont use checkpoint in line 111
#23
BAILUXIN
closed
2 years ago
0
when training, here happens 'Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to'
#22
ucasyjz
opened
2 years ago
1
SimMIM with Absolute Position Embedding
#21
hemangchawla
opened
2 years ago
0
can not reproduce your results. trained from your released pre-trained vit-base model
#20
leoozy
opened
2 years ago
1
Geometric interpolation ViT
#19
hemangchawla
opened
2 years ago
1
Plan to implement downstream tasks
#18
youngwanLEE
opened
2 years ago
0
192x192 pretraining resolution
#17
christophschuhmann
opened
2 years ago
0
Inconsistency of ViT-base config as described in the paper
#16
Jiahao000
opened
2 years ago
0
Loss goes nan after 14 epochs
#15
DianCh
opened
2 years ago
1
linear probe
#14
aaronsarna
opened
2 years ago
7
Allow arbitrary-sized images by dynamic masking: upstream changes from Swin-Transformer-Object-Detection / SOLQ
#13
vadimkantorov
opened
2 years ago
2
How can you resolve the mismatch of patch_size in <patch_embed> module between pretrained model and finetuned model?
#12
hao-pt
closed
2 years ago
2
Confusion about fine-tune
#11
Breeze-Zero
closed
2 years ago
8
why masking embedding feature maps but not input images
#10
baibaidj
closed
2 years ago
5
Could you please release the mask prediction visualization code?
#9
maxuetao
closed
2 years ago
1
USE_RPB and USE_SHARED_RPB are inconsistent in pretrain and finetune
#8
haooooooqi
closed
2 years ago
1
Could you please release the finetune log of swin and vit?
#7
zws98
opened
2 years ago
0
Can you release training log for ViT 800 epochs?
#6
gaopengcuhk
opened
2 years ago
1
About data augment
#5
peiyingxin
closed
2 years ago
1
Information about relative positional encoding
#4
marc345
closed
2 years ago
1
Missing hyperparameter
#3
haofanwang
closed
2 years ago
1
A question about masking strategy
#2
xiaohu2015
closed
2 years ago
1
When will you release the code for pre-training
#1
BaohaoLiao
closed
2 years ago
1