{"id":13564907,"url":"https://github.com/ruiminshen/yolo2-pytorch","last_synced_at":"2025-04-03T22:30:27.023Z","repository":{"id":93862679,"uuid":"115170375","full_name":"ruiminshen/yolo2-pytorch","owner":"ruiminshen","description":"PyTorch implementation of the YOLO (You Only Look Once) v2","archived":false,"fork":false,"pushed_at":"2018-05-12T13:43:37.000Z","size":34733,"stargazers_count":438,"open_issues_count":12,"forks_count":103,"subscribers_count":26,"default_branch":"master","last_synced_at":"2024-11-04T18:45:34.605Z","etag":null,"topics":["caffe2","computer-vision","deep-learning","deep-neural-networks","object-detection","onnx","python","pytorch","yolo2"],"latest_commit_sha":null,"homepage":null,"language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"lgpl-3.0","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/ruiminshen.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":"LICENSE.md","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null}},"created_at":"2017-12-23T04:30:38.000Z","updated_at":"2024-09-18T09:10:21.000Z","dependencies_parsed_at":"2023-03-29T20:49:10.072Z","dependency_job_id":null,"html_url":"https://github.com/ruiminshen/yolo2-pytorch","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/ruiminshen%2Fyolo2-pytorch","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/ruiminshen%2Fyolo2-pytorch/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/ruiminshen%2Fyolo2-pytorch/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/ruiminshen%2Fyolo2-pytorch/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/ruiminshen","download_url":"https://codeload.github.com/ruiminshen/yolo2-pytorch/tar.gz/refs/heads/master","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":247089595,"owners_count":20881802,"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":["caffe2","computer-vision","deep-learning","deep-neural-networks","object-detection","onnx","python","pytorch","yolo2"],"created_at":"2024-08-01T13:01:37.803Z","updated_at":"2025-04-03T22:30:22.013Z","avatar_url":"https://github.com/ruiminshen.png","language":"Python","funding_links":["https://www.paypal.me/minimumshen"],"categories":["Python","Other Versions of YOLO","Paper implementations｜论文实现","Paper implementations"],"sub_categories":["Other libraries｜其他库:","Other libraries:"],"readme":"# PyTorch implementation of the [YOLO (You Only Look Once) v2](https://arxiv.org/pdf/1612.08242.pdf)\n\nThe YOLOv2 is one of the most popular [one-stage](https://arxiv.org/abs/1708.02002) object detector.\nThis project adopts [PyTorch](http://pytorch.org/) as the developing framework to increase productivity, and utilize [ONNX](https://github.com/onnx/onnx) to convert models into [Caffe 2](https://caffe2.ai/) to benefit engineering deployment.\nIf you are benefited from this project, a donation will be appreciated (via [PayPal](https://www.paypal.me/minimumshen), [微信支付](donate_mm.jpg) or [支付宝](donate_alipay.jpg)).\n\n![](demo.gif)\n\n## Designs\n\n- Flexible configuration design.\nProgram settings are configurable and can be modified (via **configure file overlaping** (-c/--config option) or **command editing** (-m/--modify option)) using command line argument.\n\n- Monitoring via [TensorBoard](https://github.com/tensorflow/tensorboard).\nSuch as the loss values and the debugging images (such as IoU heatmap, ground truth and predict bounding boxes).\n\n- Parallel model training design.\nDifferent models are saved into different directories so that can be trained simultaneously.\n\n- Using a NoSQL database to store evaluation results with multiple dimension of information.\nThis design is useful when analyzing a large amount of experiment results.\n\n- Time-based output design.\nRunning information (such as the model, the summaries (produced by TensorBoard), and the evaluation results) are saved periodically via a predefined time.\n\n- Checkpoint management.\nSeveral latest checkpoint files (.pth) are preserved in the model directory and the older ones are deleted.\n\n- NaN debug.\nWhen a NaN loss is detected, the running environment (data batch) and the model will be exported to analyze the reason.\n\n- Unified data cache design.\nVarious dataset are converted into a unified data cache via corresponding cache plugins.\nSome plugins are already implemented. Such as [PASCAL VOC](http://host.robots.ox.ac.uk/pascal/VOC/) and [MS COCO](http://cocodataset.org/).\n\n- Arbitrarily replaceable model plugin design.\nThe main deep neural network (DNN) can be easily replaced via configuration settings.\nMultiple models are already provided. Such as Darknet, [ResNet](https://arxiv.org/abs/1512.03385), Inception [v3](https://arxiv.org/abs/1512.00567) and [v4](https://arxiv.org/abs/1602.07261), [MobileNet](https://arxiv.org/abs/1704.04861) and [DenseNet](https://arxiv.org/abs/1608.06993).\n\n- Extendable data preprocess plugin design.\nThe original images (in different sizes) and labels are processed via a sequence of operations to form a training batch (images with the same size, and bounding boxes list are padded).\nMultiple preprocess plugins are already implemented. Such as\naugmentation operators to process images and labels (such as random rotate and random flip) simultaneously,\noperators to resize both images and labels into a fixed size in a batch (such as random crop),\nand operators to augment images without labels (such as random blur, random saturation and random brightness).\n\n## Feautures\n\n- [x] Reproduce the original paper's training results.\n- [x] Multi-scale training.\n- [x] Dimension cluster.\n- [x] [Darknet](http://pjreddie.com) model file (`.weights`) parser.\n- [x] Detection from image and camera.\n- [x] Processing Video file.\n- [x] Multi-GPU supporting.\n- [ ] Distributed training.\n- [ ] [Focal loss](https://arxiv.org/abs/1708.02002).\n- [x] Channel-wise model parameter analyzer.\n- [x] Automatically change the number of channels.\n- [x] Receptive field analyzer.\n\n## Quick Start\n\nThis project uses [Python 3](https://www.python.org/). To install the dependent libraries, type the following command in a terminal.\n\n```\nsudo pip3 install -r requirements.txt\n```\n\n`quick_start.sh` contains the examples to perform detection and evaluation. Run this script.\nMultiple datasets and models (the original Darknet's format, will be converted into PyTorch's format) will be downloaded ([aria2](https://aria2.github.io/) is required).\nThese datasets are cached into different data profiles, and the models are evaluated over the cached data.\nThe models are used to detect objects in an example image, and the detection results will be shown.\n\n## License\n\nThis project is released as the open source software with the GNU Lesser General Public License version 3 ([LGPL v3](http://www.gnu.org/licenses/lgpl-3.0.html)).\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fruiminshen%2Fyolo2-pytorch","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fruiminshen%2Fyolo2-pytorch","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fruiminshen%2Fyolo2-pytorch/lists"}