CoinCheung

pytorch-loss. label-smooth, amsoftmax, partial-fc, focal-loss, triplet-loss, lovasz-softmax. Maybe useful

BiSeNet. Add bisenetv2. My implementation of BiSeNet

DeepLab-v3-plus-cityscapes. mIOU=80.02 on cityscapes. My implementation of deeplabv3+ (also know as 'Encoder-Decoder with Atrous Separable Convolution for Semantic Image Segmentation' based on the dataset of cityscapes).

triplet-reid-pytorch. My implementation of the paper [In Defense of the Triplet Loss for Person Re-Identification]

gdGPT. Train llm (bloom, llama, baichuan2-7b, chatglm3-6b) with deepspeed pipeline mode. Faster than zero/zero++/fsdp.

SphereReID. My implementation of paper: SphereReID: Deep Hypersphere Manifold Embedding for Person Re-Identification

fixmatch-pytorch. 90%+ with 40 labels. please see the readme for details.

Deeplab-Large-FOV. My Implementation of the deeplab_v1 (known as deeplab large fov)

MFM. code for paper "Masked Frequency Modeling for Self-Supervised Visual Pre-Training" (https://arxiv.org/pdf/2206.07706.pdf)

SFT-ReID. My implementation of Spectral-Feature-Transformation-ReID, link to the paper: https://arxiv.org/abs/1811.11405

DenseCL. DenseCL + regionCL-D

SelfSup. ssl method pretrain experiments and weights: mocov2 + fast-moco + regioncl + mixup + densecl

AutoAugment_opencv. autoaugment and randomaugment implemented with python opencv(cv2)

porn_identification. implement image processing with a combination of C and python3 based on socket communication

ImageNet-Loader. An imagenet dataloader implemented with opencv and pybind11 based on multi-threading

mixMatch. Pytorch reproduction of the experiment in the paper of MixMatch.

Segmentatron. My implementation of some segmentation algorithms

image-dedup. Codebase I use for deduplication of image datasets

GAN-mxnet. GAN implemented with mxnet

CAPTCHA_IND. Identify CAPTCHA basing on CNN implemented by mxnet

maskrcnn-benchmark. Fast, modular reference implementation of Instance Segmentation and Object Detection algorithms in PyTorch.