Codelink v2 code#
Codelink v2 series#

Codelink v2 code#

The code will be released at GitHub Repo. Swin Transformer received ICCV 2021 best paper award (Marr Prize).Integrated into Huggingface Spaces 🤗 using Gradio.SimMIM is a self-supervised pre-training approach based on masked image modeling, a key technique that works out the 3-billion-parameter Swin V2 model using 40x less labelled data than that of previous billion-scale models based on JFT-3B. Swin Transformer V2 and SimMIM got accepted by CVPR 2022.ImageNet-22K pretrained models for Swin-V1-Tiny and Swin-V2-Small are released.Pretrained models of Swin Transformer V2 on ImageNet-1K and ImageNet-22K are released.Swin-MoE is introduced in the TuTel paper. Added Swin-MoE, the Mixture-of-Experts variant of Swin Transformer implemented using Tutel (an optimized Mixture-of-Experts implementation).Merged a PR from Nvidia that enables an option to use pure FP16 (Apex O2) in training, while almost maintaining the accuracy.Merged a PR from Nvidia that links to faster Swin Transformer inference that have significant speed improvements on T4 and A100 GPUs.Particularly, it improves CLIP pre-trained ViT-L by +1.6% to reach 89.0% on ImageNet-1K image classification, which is the most accurate ViT-L model. FD is an approach that can generally improve the fine-tuning performance of various pre-trained models, including DeiT, DINO, and CLIP. SwinV2-G achieves 61.4 mIoU on ADE20K semantic segmentation (+1.5 mIoU over the previous SwinV2-G model), using an additional feature distillation (FD) approach, setting a new recrod on this benchmark.Please look into the data scaling paper for more details.

You may leverage these models to study the properties of MIM methods.

Codelink v2 series#

Released a series of Swin and SwinV2 models pre-trained using the SimMIM approach (see MODELHUB for SimMIM), with model size ranging from SwinV2-Small-50M to SwinV2-giant-1B, data size ranging from ImageNet-1K-10% to ImageNet-22K, and iterations from 125k to 500k. Please refer to get started with SimMIM to play with SimMIM pre-training. Merged SimMIM, which is a Masked Image Modeling based pre-training approach applicable to Swin and SwinV2 (and also applicable for ViT and ResNet). Please refer to Feature-Distillation for details, and the checkpoints (FD-EsViT-Swin-B, FD-DeiT-ViT-B, FD-DINO-ViT-B, FD-CLIP-ViT-B, FD-CLIP-ViT-L).

Models and codes of Feature Distillation are released.

Nvidia's FasterTransformer now supports Swin Transformer V2 inference, which have significant speed improvements on T4 and A100 GPUs.

Mixture-of-Experts: See get_started for more instructions.įeature-Distillation: See Feature-Distillation. SSL: Masked Image Modeling: See get_started.md#simmim-support. SSL: Contrasitive Learning: See Transformer-SSL. Semi-Supervised Object Detection: See Soft Teacher. Video Action Recognition: See Video Swin Transformer. Semantic Segmentation: See Swin Transformer for Semantic Segmentation. Object Detection and Instance Segmentation: See Swin Transformer for Object Detection. Image Classification: Included in this repo. It currently includes code and models for the following tasks: This repo is the official implementation of 'Swin Transformer: Hierarchical Vision Transformer using Shifted Windows' as well as the follow-ups.