README.md

SemiVisBooster

Our algorithm implementation is at semilearn/algorithms/freematch/text_match.py

TODO

• Clean the code
• Update the training

Prerequisites

Our Method is based on USB. USB is built on pytorch, with torchvision, torchaudio, and transformers.

To install the required packages, you can create a conda environment:

conda create --name usb python=3.8

then use pip to install required packages:

pip install -r requirements.txt

Prepare Datasets

The detailed instructions for downloading and processing are shown in Dataset Download. Please follow it to download datasets before running or developing algorithms.

Training

Run experiment on Food101 dataset with 404 labeled images.

python train.py --c config/usb_cv/clipfreematch/textmatch_food101_404_0_labelname_mlp.yaml

More experiment can be found under folder config/usb_cv/clipfreematch/

Evaluation

After training, you can check the evaluation performance on training logs, or running evaluation script:

python eval.py --dataset cifar100 --num_classes 100 --load_path /PATH/TO/CHECKPOINT

Acknowledgments

We thanks USB of creating our method:

• USB