{"id":13497318,"url":"https://github.com/uber-research/UPSNet","last_synced_at":"2025-03-28T21:32:22.441Z","repository":{"id":46757195,"uuid":"173386430","full_name":"uber-research/UPSNet","owner":"uber-research","description":"UPSNet: A Unified Panoptic Segmentation Network","archived":false,"fork":false,"pushed_at":"2019-07-11T21:59:54.000Z","size":168,"stargazers_count":648,"open_issues_count":76,"forks_count":119,"subscribers_count":27,"default_branch":"master","last_synced_at":"2025-03-22T06:06:09.217Z","etag":null,"topics":["computer-vision","cvpr2019","deep-learning","instance-segmentation","panoptic-segmentation","scene-parsing"],"latest_commit_sha":null,"homepage":"","language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"other","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/uber-research.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null}},"created_at":"2019-03-02T00:44:18.000Z","updated_at":"2025-01-19T09:41:56.000Z","dependencies_parsed_at":"2022-09-23T01:40:53.640Z","dependency_job_id":null,"html_url":"https://github.com/uber-research/UPSNet","commit_stats":null,"previous_names":[],"tags_count":0,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/uber-research%2FUPSNet","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/uber-research%2FUPSNet/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/uber-research%2FUPSNet/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/uber-research%2FUPSNet/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/uber-research","download_url":"https://codeload.github.com/uber-research/UPSNet/tar.gz/refs/heads/master","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":246105536,"owners_count":20724327,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2022-07-04T15:15:14.044Z","host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["computer-vision","cvpr2019","deep-learning","instance-segmentation","panoptic-segmentation","scene-parsing"],"created_at":"2024-07-31T20:00:28.791Z","updated_at":"2025-03-28T21:32:22.121Z","avatar_url":"https://github.com/uber-research.png","language":"Python","funding_links":[],"categories":["SemanticSeg"],"sub_categories":[],"readme":"# UPSNet: A Unified Panoptic Segmentation Network\n\n# Introduction\nUPSNet is initially described in a [CVPR 2019 oral](https://arxiv.org/abs/1901.03784) paper.\n\n\n\n\n# Disclaimer\n\nThis repository is tested under Python 3.6, PyTorch 0.4.1. And model training is done with 16 GPUs by using [horovod](https://github.com/horovod/horovod). It should also work under Python 2.7 / PyTorch 1.0 and with 4 GPUs.\n\n# License\n© Uber, 2018-2019. Licensed under the Uber Non-Commercial License.\n\n# Citing UPSNet\n\nIf you find UPSNet is useful in your research, please consider citing:\n```\n@inproceedings{xiong19upsnet,\n    Author = {Yuwen Xiong, Renjie Liao, Hengshuang Zhao, Rui Hu, Min Bai, Ersin Yumer, Raquel Urtasun},\n    Title = {UPSNet: A Unified Panoptic Segmentation Network},\n    Conference = {CVPR},\n    Year = {2019}\n}\n```\n\n\n# Main Results\n\nCOCO 2017 (trained on train-2017 set)\n\n|                | test split | PQ   | SQ   | RQ   | PQ\u003csup\u003eTh\u003c/sup\u003e | PQ\u003csup\u003eSt\u003c/sup\u003e |\n|----------------|------------|------|------|------|-----------------|-----------------|\n| UPSNet-50      | val        | 42.5 | 78.0 | 52.4 | 48.5            | 33.4            |\n| UPSNet-101-DCN | test-dev   | 46.6 | 80.5 | 56.9 | 53.2            | 36.7            |\n\nCityscapes\n\n|                | PQ   | SQ   | RQ   | PQ\u003csup\u003eTh\u003c/sup\u003e | PQ\u003csup\u003eSt\u003c/sup\u003e |\n|----------------|------|------|------|-----------------|-----------------|\n| UPSNet-50      | 59.3 | 79.7 | 73.0 | 54.6            | 62.7            |\n| UPSNet-101-COCO (ms test) | 61.8 | 81.3 | 74.8 | 57.6 | 64.8 |\n\n# Requirements: Software\n\nWe recommend using Anaconda3 as it already includes many common packages.\n\n\n# Requirements: Hardware\n\nWe recommend using 4~16 GPUs with at least 11 GB memory to train our model.\n\n# Installation\n\nClone this repo to `$UPSNet_ROOT`\n\nRun `init.sh` to build essential C++/CUDA modules and download pretrained model.\n\nFor Cityscapes:\n\nAssuming you already downloaded Cityscapes dataset at `$CITYSCAPES_ROOT` and TrainIds label images are generated, please create a soft link by `ln -s $CITYSCAPES_ROOT data/cityscapes` under `UPSNet_ROOT`, and run `init_cityscapes.sh` to prepare Cityscapes dataset for UPSNet.\n\nFor COCO:\n\nAssuming you already downloaded COCO dataset at `$COCO_ROOT` and have `annotations` and `images` folders under it, please create a soft link by `ln -s $COCO_ROOT data/coco` under `UPSNet_ROOT`, and run `init_coco.sh` to prepare COCO dataset for UPSNet.\n\nTraining:\n\n`python upsnet/upsnet_end2end_train.py --cfg upsnet/experiments/$EXP.yaml`\n\nTest:\n\n`python upsnet/upsnet_end2end_test.py --cfg upsnet/experiments/$EXP.yaml`\n\nWe provide serveral config files (16/4 GPUs for Cityscapes/COCO dataset) under upsnet/experiments folder.\n\n# Model Weights\n\nThe model weights that can reproduce numbers in our paper are available now. Please follow these steps to use them:\n\nRun `download_weights.sh` to get trained model weights for Cityscapes and COCO.\n\nFor Cityscapes:\n\n```shell\npython upsnet/upsnet_end2end_test.py --cfg upsnet/experiments/upsnet_resnet50_cityscapes_16gpu.yaml --weight_path ./model/upsnet_resnet_50_cityscapes_12000.pth\n```\n\n```shell\npython upsnet/upsnet_end2end_test.py --cfg upsnet/experiments/upsnet_resnet101_cityscapes_w_coco_16gpu.yaml --weight_path ./model/upsnet_resnet_101_cityscapes_w_coco_3000.pth\n```\n\nFor COCO:\n\n```shell\npython upsnet/upsnet_end2end_test.py --cfg upsnet/experiments/upsnet_resnet50_coco_16gpu.yaml --weight_path model/upsnet_resnet_50_coco_90000.pth\n```\n\n```shell\npython upsnet/upsnet_end2end_test.py --cfg upsnet/experiments/upsnet_resnet101_dcn_coco_3x_16gpu.yaml --weight_path model/upsnet_resnet_101_dcn_coco_270000.pth\n```\n\n\n\n\n\n\n\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fuber-research%2FUPSNet","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fuber-research%2FUPSNet","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fuber-research%2FUPSNet/lists"}