Comments (2)
Yeah we also experienced something similar. It can be because of numerous reasons:
- When you're changing the pretraining dataset usually other hyperparameters need to be changed as well. One of the most important component is the augmentation pipeline especially the crop ratio for random crops, strength of color distortion, and so on. We didn't run extensive experiments with these
- You can refer to these guidelines from the original SwAV authors: https://github.com/facebookresearch/swav#common-issues. We tried some of these and got better results but didn't have the time to update things here. Sorry about that.
from swav-tf.
from swav-tf.
Related Issues (9)
- Pointers on Train_Step_And_Loss.ipynb HOT 6
- Copyright information HOT 1
- image loaders don't produce augmentations from the same image HOT 4
- Classify a single image HOT 2
- Pointers on Minimal_Sinkhorn_Knopp.ipynb HOT 1
- Fine-tuning with 10% labeled data with SwAV-learned embeddings
- Linear evaluation on the entire dataset with pre-trained ResNet50 features
- [QUESTION] Difference 10-epochs, 40-epochs HOT 3
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from swav-tf.