Argilla is a collaboration tool for AI engineers and domain experts to build high-quality datasets

Code for ALBEF: a new vision-language pre-training method

Open source tools for computational pathology - Nature BME

Cross-Domain Weakly-Supervised Object Detection through Progressive Domain Adaptation [Inoue+, CVPR2018].

DSMIL: Dual-stream multiple instance learning networks for tumor detection in Whole Slide Image

Single-Stage Semantic Segmentation from Image Labels (CVPR 2020)

Weakly-supervised object detection.

Mask-Free Video Instance Segmentation [CVPR 2023]

Weakly Supervised Learning for Findings Detection in Medical Images

[EMNLP 2020] Text Classification Using Label Names Only: A Language Model Self-Training Approach

BOND: BERT-Assisted Open-Domain Name Entity Recognition with Distant Supervision

PyTorch implementation of "WILDCAT: Weakly Supervised Learning of Deep ConvNets for Image Classification, Pointwise Localization and Segmentation", CVPR 2017

[NeurIPS 2021] WRENCH: Weak supeRvision bENCHmark

Scribbles or Points-based weakly-supervised learning for medical image segmentation, a strong baseline, and tutorial for research and application.

Caffe codes for our papers "Multiple Instance Detection Network with Online Instance Classifier Refinement" and "PCL: Proposal Cluster Learning for Weakly Supervised Object Detection".

[NAACL 2021] This is the code for our paper `Fine-Tuning Pre-trained Language Model with Weak Supervision: A Contrastive-Regularized Self-Training Approach'.

Attention-based Dropout Layer for Weakly Supervised Object Localization (CVPR 2019 Oral)

[CVPR 2022] C2AM: Contrastive learning of Class-agnostic Activation Map for Weakly Supervised Object Localization and Semantic Segmentation

[ICIP 2021] Puzzle-CAM: Improved localization via matching partial and full features.

Official Pytorch Implementation of 'Background Suppression Network for Weakly-supervised Temporal Action Localization' (AAAI-20 Spotlight)