PyTorch Implementation: FixMatch

PyTorch implementation of FixMatch: Simplifying Semi-Supervised Learning with Consistency and Confidence based on the official tensorflow implementation.

The implementation supports the following datasets:

• CIFAR-10 / CIFAR-100
• SVHN
• Caltech101 / Caltech256
• STL10
• HAM10000
• ImageNet

Installation

Required python packages are listed in requirements.txt. All dependencies can be installed using pip

pip install -r requirements.txt

or using conda

conda install --file requirements.txt

Training

FixMatch training is started by running the following command (--pbar to show progress bar during training):

python main.py --pbar

All commandline arguments, which can be used to adapt the configuration of FixMatch are defined and described in arguments.py. By default the following configuration is run:

model: 'wide_resnet28_2'
dataset: 'cifar10'
lr: 0.03
wd: 0.0005
num_labeled: 250 (number of labeled samples, i.e. 25 labeled samples per class for cifar10)
iters_per_epoch: 1024
batch_size: 64
epochs: 1024
device: 'cuda'
out_dir: 'fixmatch'
m: 30
threshold: 0.95
beta: 0.9
num_augmentations: 2
mu: 7
wu: 1

In addition to these, the following arguments can be used to further configure the FixMatch training process:

• --device <cuda / cpu>: Specify whether training should be run on GPU (if available) or CPU
• --num-workers <num_workers>: Number of workers used by torch dataloader
• --resume <path to run_folder>: Resumes training of training run saved at specified path, e.g. 'out/mixmatch_training/run_0'. Dataset splits, model state, optimizer state, etc. are loaded and training is resumed with specified arguments.
• see arguments.py for more

Alternatively, the polyaxon.yaml-file can be used to start FixMatch training on a polyaxon-cluster:

polyaxon run -f polyaxon.yaml -u

For a general introduction to polyaxon and its commandline client, please refer to the official documentation

Monitoring

The training progress (loss, accuracy, etc.) can be monitored using tensorboard as follows:

tensorboard --logdir <result_folder>

This starts a tensorboard instance at localhost:6006, which can be opened in any common browser.

Evaluation

A trained FixMatch model can be evaluated by running:

 python3 eval.py --run-path out/fixmatch_training/run_0 --pbar --device <cuda / cpu>

where --run-path specifies the path at which the run to be evaluated is saved. Alternatively, one can also check all metrics over all epochs using the tensorboard file.

References

@article{sohn2020fixmatch,
    title={FixMatch: Simplifying Semi-Supervised Learning with Consistency and Confidence},
    author={Kihyuk Sohn and David Berthelot and Chun-Liang Li and Zizhao Zhang and Nicholas Carlini and Ekin D. Cubuk and Alex Kurakin and Han Zhang and Colin Raffel},
    journal={arXiv preprint arXiv:2001.07685},
    year={2020},
}