Comments (5)
Hi @prigoyal do you need help with the BYOL baseline
from vissl.
@prigoyal ok. I need time to understand your code first. Currently, I am working on BYOL with another idea, so it seems I can contribute the baseline to your framework once I understand your code well.
from vissl.
Hi @chaoyanghe , thank you so much for reaching out. Yes , absolutely would love help. Feel free to grab the issue and assign to yourself (or someone) :)
Highly appreciated and the contribution will be very impactful for the SSL community!!
from vissl.
@chaoyanghe that sounds great. Just sharing some pointers that could be helpful:
In the VISSL docs, I created several flowcharts https://vissl.readthedocs.io/en/latest/flowcharts/train_workflow.html that could be helpful to understand the execution.
Further look at how to extend several components of VISSL in the docs https://vissl.readthedocs.io/en/latest/extend_modules/data_transforms.html
My understanding with BYOL is that : we would need to ensure that the data transforms are correct + make sure the loss is correct , the head of the model is correct -> all these components are modular and the docs above will be helpful.
Also look at the MoCo hook https://github.com/facebookresearch/vissl/blob/master/vissl/hooks/moco_hooks.py and moco loss https://github.com/facebookresearch/vissl/blob/master/vissl/losses/moco_loss.py for how we handled the momentum encoder
Lastly, feel free to reach out on this issue anytime :) thank you again! :)
from vissl.
cc @iseessel
from vissl.
Related Issues (20)
- How to run the benchmark obeject detection task on VOC dataset? HOT 1
- Linear Evaluation on VOC07 HOT 1
- Unsupervised pre-training on custom data HOT 2
- Example of using ZeRO and LARC
- Incosistent results when running inference on trained resnet model HOT 1
- Low performance of Supervised training
- MLP_DIM in ViT is not adjustable through the config. HOT 2
- [Feature] Online linear probing
- Linear Image classification weights not loading properly: vissl pretrained models simclr or Dcv2 HOT 3
- "on_forward"-hook operations and back-propagation
- EMA does not work on fp16 and does not copy weights?
- First-class support for timm models
- Can't load pretrained seer/swav checkpoints (RG_Y_128GF) HOT 1
- Is it possible to get full imagenet pretrained weights of MoCo v2 ? HOT 2
- NPID + MoCoV2 weights are the same? HOT 1
- Potential confusion over licensing comment
- Does the MoCo implementation do shuffled Batch Norm when run on a single GPU? HOT 3
- Loading TRAINED models with VISSL HOT 9
- Use pretrain outside of vissl
- can someone share/has anyone shared a checkpoint ( with the class head) of a ssl model on imagenet?
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 vissl.