Implementation of Mean Teacher

This Repo is the implementation of the following paper

• Mean Teacher Mean teachers are better role models: Weight-averaged consistency targets improve semi-supervised deep learning results

We used the Convlarge architecture to train Mean Teacher

checkpoints are set after each epoch to save models and parameters. Ready for plot.

versions we use:

1. Pytorch 1.6.0
2. Python 3.7.3
3. torchvision 0.7.0
4. cudatoolkit 10.2
5. TensorboardX

also work: conda install pytorch==1.7.1 torchvision==0.8.2 torchaudio==0.7.2 cudatoolkit=11.0 -c pytorch

Dataset

we didn't include dataset, but after runing the following shell, three folders under "...\data-local\images\cifar\cifar10\by-image" should be \train, \test, and \val

./data-local/bin/prepare_cifar10.sh

Architecture

Accuracy Achieved on Test Dataset

A step by step series of examples that tell you how to get a development env running

Say what the step will be

Mean Teacher
    a) Student Model : 79%
    b) Teacher Model : 80%

Running the Training

This repo implemented several methods and control them using flags.

For Mean Teacher

Go the parameters.py and change the following flags as follows:

1. supervised_mode = False ( To use only 4000 labels for training)
2. lr = 0.2 ( setting the learning rate)
3. BN = False or True ( for turning batch Normalization on or off)
4. sntg = False ( Do not use any SNTG loss )
5. Do not change any other settings and run main.py

Tensorboard Visualization

To Visualize on Tensorboard, use the following command

tensorboard --logdir=”path to ./ckpt”

Note that all the checkpoints are in the ./ckpt folder so simply start a tensorboard session to visualize it. Also all the saved checkpoints for student models are also saved there.

1. Baseline : 12-03-18:09/convlarge,Adam,200epochs,b256,lr0.15/test
2. Mean teacher without BN :
   12-03-20:12/convlarge,Adam,200epochs,b256,lr0.15/test
   12-03-23:38/convlarge,Adam,200epochs,b256,lr0.2/test
3. Mean Teacher with BN : 12-05-11:55/convlarge,Adam,200epochs,b256,lr0.2/test
4. Hybrid Net : 12-06-10:58/hybridnet,Adam,200epochs,b256,lr0.2/test
5. SNTG + Meant Teacher: 12-07-00:36/convlarge,Adam,200epochs,b256,lr0.2/test

Acknowledgments

This implementation has been modified based on the following source.

• Mean Teacher : We have mainly followed this Repo, but did necessary modification to make the code run on Python 3.7.x and the visualize the results graphically.