facebookresearch/msn
Masked Siamese Networks for Label-Efficient Learning (https://arxiv.org/abs/2204.07141)
PythonNOASSERTION
Issues
- 7
pretrained weights are probably incorrect
#7 opened by BestSonny - 4
Duplicate GPU detected error
#24 opened by amlivne - 1
No such file
#23 opened by chaichaichai - 0
The loss has converged at early stage?
#22 opened by sinsauzero - 1
How to change MSN loss to PMSN loss? (from paper "The Hidden Uniform Cluster Prior in Self-Supervised Learning")
#20 opened by faris-k - 0
What is the use of "AllReduce"?
#21 opened by yyk-wew - 0
Using just the encoder
#19 opened by danik7 - 1
Include full checkpoint
#12 opened by mhamilton723 - 0
- 1
why did not take block-wise mask strategy?
#16 opened by ShiYaya - 0
vit-b-16 config
#17 opened by geekJZY - 0
Performance when small batch size
#15 opened by SongDoHou - 1
The detail setting for 1% evaluation
#11 opened by Haoqing-Wang - 0
Lambda for logistic regression evaluation
#14 opened by harihp12 - 0
Custom dataset Linear Eval
#13 opened by letdivedeep - 1
About the 1% In1k semi-sup evaluation
#9 opened by merlinarer - 4
- 1
- 1
linear_eval.py #L327
#6 opened by Lee-YZz - 1
- 2
Why we need multiply 1.25 for num_epochs?
#4 opened by andy-yangz - 1
add web demo/model to Huggingface
#1 opened by AK391 - 1
Segmentation Performance
#2 opened by cyrilzakka - 0