SimMatchV2: Semi-Supervised Learning with Graph Consistency
Semi-Supervised image classification is one of the most fundamental problem in computer vision, which significantly reduces the need for human labor.
Tags：Paper and LLMsNode Classification Semi-Supervised Image ClassificationPricing Type
- Pricing Type: Free
- Price Range Start($):
GitHub Link
The GitHub link is https://github.com/mingkai-zheng/simmatchv2
Introduce
The GitHub repository “SimMatchV2” presents code and pretrained models for the ICCV2023 paper on “SimMatchV2 Semi-Supervised Learning with Graph Consistency.” The repository includes PyTorch evaluation and training code. The code is adapted from a reference source and is designed for distributed training in a slurm environment. Pretrained models for SimMatchV2 with different settings are provided, along with their corresponding accuracy percentages. Researchers can also find evaluation scripts and citation details in the repository. The paper’s authors are Zheng, Mingkai et al., and the work focuses on semi-supervised learning using graph consistency.
Semi-Supervised image classification is one of the most fundamental problem in computer vision, which significantly reduces the need for human labor.
Content
This repository contains PyTorch evaluation code, training code and pretrained models for SimMatchV2. Most of the code in this repository is adapt from here. For details see SimMatchV2: Semi-Supervised Learning with Graph Consistency by Mingkai Zheng, Shan You, Lang Huang, Chen Luo, Fei Wang, Chen Qian, and Chang Xu To run the code, you probably need to change the Dataset setting (ImagenetPercent function in data/imagenet.py), and Pytorch DDP setting (dist_init function in utils/dist_utils.py) for your server environment. The distributed training of this code is based on slurm environment, we have provided the training scrips in script/train.sh We also provide the pre-trained model. If you want to test the pre-trained model, please download the weights from the link above, and move them to the checkpoints folder. The evaluation scripts also have been provided in script/train.sh If you find that SimMatch interesting and help your research, please consider citing it:
Related
LAMA utilizes a reinforcement learning framework combined with a motion matching algorithm. Reinforcement learning helps the model make appropriate decisions in various scenarios, while motion matching algorithms ensure that synthesized actions match real human actions. In addition, LAMA also utilizes the motion editing framework of manifold learning to cover various possible changes in interactions and operations.
