{"id":16568408,"url":"https://github.com/breezedeus/loveshare","last_synced_at":"2026-03-06T09:31:20.975Z","repository":{"id":87336345,"uuid":"93234867","full_name":"breezedeus/LoveShare","owner":"breezedeus","description":"breezedeus的各种分享","archived":false,"fork":false,"pushed_at":"2023-01-31T13:39:13.000Z","size":41549,"stargazers_count":22,"open_issues_count":0,"forks_count":5,"subscribers_count":3,"default_branch":"master","last_synced_at":"2025-03-05T11:40:44.449Z","etag":null,"topics":["cnocr","cnstd","cv","deep-learning","llm","nlp","ocr","pix2text"],"latest_commit_sha":null,"homepage":"https://www.breezedeus.com","language":"TeX","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":null,"status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/breezedeus.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":null,"code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null}},"created_at":"2017-06-03T08:33:46.000Z","updated_at":"2025-02-21T07:22:26.000Z","dependencies_parsed_at":null,"dependency_job_id":"d03a236d-db28-4cd9-9270-5a41982c92b0","html_url":"https://github.com/breezedeus/LoveShare","commit_stats":null,"previous_names":[],"tags_count":1,"template":false,"template_full_name":null,"purl":"pkg:github/breezedeus/LoveShare","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/breezedeus%2FLoveShare","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/breezedeus%2FLoveShare/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/breezedeus%2FLoveShare/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/breezedeus%2FLoveShare/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/breezedeus","download_url":"https://codeload.github.com/breezedeus/LoveShare/tar.gz/refs/heads/master","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/breezedeus%2FLoveShare/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":286080680,"owners_count":30168969,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2026-03-06T07:56:45.623Z","status":"ssl_error","status_checked_at":"2026-03-06T07:55:55.621Z","response_time":250,"last_error":"SSL_connect returned=1 errno=0 peeraddr=140.82.121.6:443 state=error: unexpected eof while reading","robots_txt_status":"success","robots_txt_updated_at":"2025-07-24T06:49:26.215Z","robots_txt_url":"https://github.com/robots.txt","online":false,"can_crawl_api":true,"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":["cnocr","cnstd","cv","deep-learning","llm","nlp","ocr","pix2text"],"created_at":"2024-10-11T21:09:24.634Z","updated_at":"2026-03-06T09:31:20.666Z","avatar_url":"https://github.com/breezedeus.png","language":"TeX","funding_links":[],"categories":[],"sub_categories":[],"readme":"# Year 2022\n\n* **[视觉-语言预训练（VLP）技术介绍](https://www.bilibili.com/video/BV1dr4y1E7ZR)** [![bilibili](https://img.shields.io/badge/dynamic/json?label=views\u0026style=social\u0026logo=bilibili\u0026query=data.stat.view\u0026url=https%3A%2F%2Fapi.bilibili.com%2Fx%2Fweb-interface%2Fview%3Fbvid%3DBV1dr4y1E7ZR)](https://www.bilibili.com/video/BV1dr4y1E7ZR)\n    * 分享对应的 [PPT](2022/Intro-to-VLP.pdf)；\n\n    * 对当前的视觉-语言预训练（Vision-Language Pretraining, VLP) 技术做了概要介绍，也可以看成是对当前多模态学习（Multi-Modal Learning）技术的概要介绍。\n\n* **[信息抽取统一框架介绍和实例—— UIE (Universal Information Extraction)](https://www.bilibili.com/video/BV1LW4y1U7ch)** [![bilibili](https://img.shields.io/badge/dynamic/json?label=views\u0026style=social\u0026logo=bilibili\u0026query=data.stat.view\u0026url=https%3A%2F%2Fapi.bilibili.com%2Fx%2Fweb-interface%2Fview%3Fbvid%3DBV1LW4y1U7ch)](https://www.bilibili.com/video/BV1LW4y1U7ch)\n    * 分享对应的 [文档](2022/UIE_Universal_Information_Extraction.pdf)；\n    * 介绍了百度 ACL 2022 的论文《Unified Structure Generation for Universal Information Extraction》，并基于百度开源的代码跑了训练示例。\n\n* **[BERT 还有哪些神奇的能力呢！？ - 知乎](https://zhuanlan.zhihu.com/p/532010499)**\n\n\n\n### 开源工具的使用介绍\n\n* **[Pix2Text: 替代 Mathpix 的免费 Python 开源工具](https://www.bilibili.com/video/BV12e4y1871U)** [![bilibili](https://img.shields.io/badge/dynamic/json?label=views\u0026style=social\u0026logo=bilibili\u0026query=data.stat.view\u0026url=https%3A%2F%2Fapi.bilibili.com%2Fx%2Fweb-interface%2Fview%3Fbvid%3DBV12e4y1871U)](https://www.bilibili.com/video/BV12e4y1871U)\n  * **[Pix2Text (P2T)](https://github.com/breezedeus/pix2text)** ，目标是 **Mathpix** 的Python开源替代品，现在可以识别截图中的数学公式并转换为Latex表示，也可以识别图片中的中英文文字；\n  * 知乎文章（文字版）：[Pix2Text: 替代 Mathpix 的免费 Python 开源工具](https://zhuanlan.zhihu.com/p/566498651)\n  * 在线Demo： https://huggingface.co/spaces/breezedeus/pix2text ；\n  * **Gitee** Fork: https://gitee.com/breezedeus/pix2text 。\n\n* **[更更好用的开源Python OCR工具包——CnOCR V2.2 - 知乎](https://zhuanlan.zhihu.com/p/546441117)**\n  * 在线Demo： https://huggingface.co/spaces/breezedeus/cnocr 。\n* **[利用CnOCR实现自动对截屏图片OCR - 知乎](https://zhuanlan.zhihu.com/p/554108141)**\n* **[如何安装CnOCR，以及免安装直接使用CnOCR](https://www.bilibili.com/video/BV1NY4y1T7jG)** [![bilibili](https://img.shields.io/badge/dynamic/json?label=views\u0026style=social\u0026logo=bilibili\u0026query=data.stat.view\u0026url=https%3A%2F%2Fapi.bilibili.com%2Fx%2Fweb-interface%2Fview%3Fbvid%3DBV1NY4y1T7jG)](https://www.bilibili.com/video/BV1NY4y1T7jG)\n  * 分享对应的 [PPT](2022/CnOCR-Installation-FAQ.pdf)；\n  * 介绍了 **[CnOCR](https://github.com/breezedeus/cnocr)** 安装和使用过程中大家遇到的一些问题。能不能不安装试用CnOCR效果？没问题，本视频中提供了**4种免安装**使用的方法。\n  * 相关资源：[文本检测和识别——附CnStd与CnOcr工具介绍](https://www.bilibili.com/video/BV1uU4y1N7Ba) [![bilibili](https://img.shields.io/badge/dynamic/json?label=views\u0026style=social\u0026logo=bilibili\u0026query=data.stat.view\u0026url=https%3A%2F%2Fapi.bilibili.com%2Fx%2Fweb-interface%2Fview%3Fbvid%3DBV1uU4y1N7Ba)](https://www.bilibili.com/video/BV1uU4y1N7Ba)\n\n* **[antiOCR 使用介绍](https://www.bilibili.com/video/BV1q8411G7HE)** [![bilibili](https://img.shields.io/badge/dynamic/json?label=views\u0026style=social\u0026logo=bilibili\u0026query=data.stat.view\u0026url=https%3A%2F%2Fapi.bilibili.com%2Fx%2Fweb-interface%2Fview%3Fbvid%3DBV1q8411G7HE)](https://www.bilibili.com/video/BV1q8411G7HE)\n  * **[antiOCR](https://github.com/breezedeus/antiOCR)** 把指定文本转换成机器无法识别但人可读的文字图片，即拒绝图片内容被OCR。常见的使用场景是图片验证码。antiOCR 是Python工具包，安装即可使用，或者直接使用作者提供的在线Demo生成图片；\n  * 在线Demo：https://huggingface.co/spaces/breezedeus/antiOCR 。\n\n\n\n# Year 2021\n\n* **[Tricks for Sparse \u0026 Dense Retrieval Models](https://www.bilibili.com/video/BV13y4y1R71Q)** [![bilibili](https://img.shields.io/badge/dynamic/json?label=views\u0026style=social\u0026logo=bilibili\u0026query=data.stat.view\u0026url=https%3A%2F%2Fapi.bilibili.com%2Fx%2Fweb-interface%2Fview%3Fbvid%3DBV13y4y1R71Q)](https://www.bilibili.com/video/BV13y4y1R71Q)\n\n  * 分享对应的 [PPT](2021/sparse-dense-retrieval.pdf)；\n  * 介绍了 **Sparse Retrieval (SR)** 和  **Dense Retrieval (DR)** 中的几大类改进方法。 *虽然是 2021年底的分享，不过到 2023年才发布出来，拖延症😅。*\n\n* **[文本检测和识别——附CnStd与CnOcr工具介绍](https://www.bilibili.com/video/BV1uU4y1N7Ba)** [![bilibili](https://img.shields.io/badge/dynamic/json?label=views\u0026style=social\u0026logo=bilibili\u0026query=data.stat.view\u0026url=https%3A%2F%2Fapi.bilibili.com%2Fx%2Fweb-interface%2Fview%3Fbvid%3DBV1uU4y1N7Ba)](https://www.bilibili.com/video/BV1uU4y1N7Ba)\n\n  * 分享对应的 [PPT](2021/cnstd-cnocr.pdf)；\n  * 介绍了文本检测和识别中的深度学习框架与代表性算法，之后介绍了 **[CnSTD](https://github.com/breezedeus/cnstd)** 与 **[CnOCR](https://github.com/breezedeus/cnocr)** 两个python3工具包的使用。\n\n* **[RS论文阅读：你真的读懂了Youtube DNN推荐论文吗？](https://www.bilibili.com/video/BV1BK4y1d7gD)** [![bilibili](https://img.shields.io/badge/dynamic/json?label=views\u0026style=social\u0026logo=bilibili\u0026query=data.stat.view\u0026url=https%3A%2F%2Fapi.bilibili.com%2Fx%2Fweb-interface%2Fview%3Fbvid%3DBV1BK4y1d7gD)](https://www.bilibili.com/video/BV1BK4y1d7gD)\n\n  * 分享的 [文字版](https://zhuanlan.zhihu.com/p/372238343) ；\n  * Youtube 在2016年发表的DNN推荐论文是最早把深度学习成功应用于推荐的代表性工作。很多博客都对这篇文章进行过翻译和解读，但你真的读懂了吗？本视频将带你温习这篇经典论文，并让你真的读懂它。\n\n* **[线上测试 I：如何快速进行线上测试](https://www.bilibili.com/video/BV18b4y1972v)** [![bilibili](https://img.shields.io/badge/dynamic/json?label=views\u0026style=social\u0026logo=bilibili\u0026query=data.stat.view\u0026url=https%3A%2F%2Fapi.bilibili.com%2Fx%2Fweb-interface%2Fview%3Fbvid%3DBV18b4y1972v)](https://www.bilibili.com/video/BV18b4y1972v) \n\n  * 介绍了线上测试的两种方法：**A/B测试** 和 **交错测试**，适用于算法测试、产品增长测试等场景。\n\n* **[线上测试 II：效果评估——假设检验](https://www.bilibili.com/video/BV1LK4y1S7uR)** [![bilibili](https://img.shields.io/badge/dynamic/json?label=views\u0026style=social\u0026logo=bilibili\u0026query=data.stat.view\u0026url=https%3A%2F%2Fapi.bilibili.com%2Fx%2Fweb-interface%2Fview%3Fbvid%3DBV1LK4y1S7uR)](https://www.bilibili.com/video/BV1LK4y1S7uR)\n\n  * 介绍了客观评估线上实验效果的数学方法：**假设检验**。\n\n* **[声纹分割聚类（Speaker Diarization）](https://www.bilibili.com/video/BV1rp4y1q7HW)** [![bilibili](https://img.shields.io/badge/dynamic/json?label=views\u0026style=social\u0026logo=bilibili\u0026query=data.stat.view\u0026url=https%3A%2F%2Fapi.bilibili.com%2Fx%2Fweb-interface%2Fview%3Fbvid%3DBV1rp4y1q7HW)](https://www.bilibili.com/video/BV1rp4y1q7HW) \n\n  * 分享的 [文字版](https://zhuanlan.zhihu.com/p/338656027) ；\n  * 介绍了声纹分割聚类的整体流程以及涉及的技术知识。\n\n* **[更好地理解图片：场景文字检测工具 cnstd - 知乎](https://zhuanlan.zhihu.com/p/145913973)**\n\n* **[2021年云栖大会见闻分享](https://www.bilibili.com/video/BV1bq4y1k7yr)** [![bilibili](https://img.shields.io/badge/dynamic/json?label=views\u0026style=social\u0026logo=bilibili\u0026query=data.stat.view\u0026url=https%3A%2F%2Fapi.bilibili.com%2Fx%2Fweb-interface%2Fview%3Fbvid%3DBV1bq4y1k7yr)](https://www.bilibili.com/video/BV1bq4y1k7yr)\n\n  * 分享在2021年云栖大会的见闻，包括视频理解、NLP、多模态、虚拟人、元宇宙等内容，无技术细节。\n\n  \n\n# Year 2020\n\n* **[如何做调研](https://www.bilibili.com/video/BV1tD4y127Xq/)** [![bilibili](https://img.shields.io/badge/dynamic/json?label=views\u0026style=social\u0026logo=bilibili\u0026query=data.stat.view\u0026url=https%3A%2F%2Fapi.bilibili.com%2Fx%2Fweb-interface%2Fview%3Fbvid%3DBV1tD4y127Xq)](https://www.bilibili.com/video/BV1tD4y127Xq)\n  * 分享对应的 [脑图](如何做调研.png)；\n  * 介绍了不同层面的调研该怎么做，包括竞品调研，针对某个具体应用的方案调研，针对某个算法（任务）方向的调研。同时介绍了一些常用的调研工具。\n\n* **[自监督学习\u0026对比学习](https://www.bilibili.com/video/BV1v5411x7rD)**  [![bilibili](https://img.shields.io/badge/dynamic/json?label=views\u0026style=social\u0026logo=bilibili\u0026query=data.stat.view\u0026url=https%3A%2F%2Fapi.bilibili.com%2Fx%2Fweb-interface%2Fview%3Fbvid%3DBV1v5411x7rD)](https://www.bilibili.com/video/BV1v5411x7rD)\n  * 分享对应的 [PPT](2020/自监督学习-对比学习.pdf)；\n  * 介绍深度学习中最前沿的自监督学习和对比学习技术。\n\n* **[NLP中的自监督学习\u0026对比学习](https://www.bilibili.com/video/BV13T4y1c73g)**  [![bilibili](https://img.shields.io/badge/dynamic/json?label=views\u0026style=social\u0026logo=bilibili\u0026query=data.stat.view\u0026url=https%3A%2F%2Fapi.bilibili.com%2Fx%2Fweb-interface%2Fview%3Fbvid%3DBV13T4y1c73g)](https://www.bilibili.com/video/BV13T4y1c73g)\n  * 分享对应的 [PPT](2020/自监督学习-对比学习2-NLP.pdf)；\n  * self-supervised \u0026 contrastive learning 第二弹：介绍NLP领域自监督学习和对比学习的最新工作（2020年）。 \n  * self-supervised \u0026 contrastive learning 第一弹：[自监督学习\u0026对比学习](https://www.bilibili.com/video/BV1v5411x7rD)  [![bilibili](https://img.shields.io/badge/dynamic/json?label=views\u0026style=social\u0026logo=bilibili\u0026query=data.stat.view\u0026url=https%3A%2F%2Fapi.bilibili.com%2Fx%2Fweb-interface%2Fview%3Fbvid%3DBV1v5411x7rD)](https://www.bilibili.com/video/BV1v5411x7rD)\n\n* **[最新语音识别技术简介（Introduction to ASR）](https://www.bilibili.com/video/BV1fZ4y1g7UP)** [![bilibili](https://img.shields.io/badge/dynamic/json?label=views\u0026style=social\u0026logo=bilibili\u0026query=data.stat.view\u0026url=https%3A%2F%2Fapi.bilibili.com%2Fx%2Fweb-interface%2Fview%3Fbvid%3DBV1fZ4y1g7UP)](https://www.bilibili.com/video/BV1fZ4y1g7UP) \n  * 分享对应的 [PPT](2020/Intro-to-ASR.pdf)；\n  * 本次分享对最新的语音识别技术做了整体的介绍。\n* **[开放域聊天机器人技术介绍](https://www.bilibili.com/video/BV1e5411Y7ci)** [![bilibili](https://img.shields.io/badge/dynamic/json?label=views\u0026style=social\u0026logo=bilibili\u0026query=data.stat.view\u0026url=https%3A%2F%2Fapi.bilibili.com%2Fx%2Fweb-interface%2Fview%3Fbvid%3DBV1e5411Y7ci)](https://www.bilibili.com/video/BV1e5411Y7ci) \n  * 分享对应的 [PPT](2020/open-domain-chatbot.pdf)；\n  * 分享的 [文字版](https://zhuanlan.zhihu.com/p/150608851) ；\n  * 主要介绍 Google Meena 和 Facebook Blender 这两个工作，它们分别发表于今年1月和4月（他们远程办公的效率看来很高）。这两篇论文在模型方面都没有什么创新，但融合了很多有意思的技术，这些模型之外的技术值得做对话的同学了解一下。现在大概是聊天机器人的 GPT-1 时代，谁会开启聊天机器人的 BERT 时代呢？\n* **[微软小冰对话机器人架构 - 知乎](https://zhuanlan.zhihu.com/p/57532328)**\n\n\n\n\n# Old Notes\n\n## Graphical Models\n\n**图模型、VEM**等相关知识的[介绍pdf](./Graphical_Models.pdf) 。\n\n\n\n## Bayes Factors\n\n**贝叶斯因子**（**Bayes Factors**）一篇论文的 [学习笔记pdf](./Bayes_Factors.pdf) 。\n\n\n\n\n\n---\n\n**官方库**：[https://github.com/breezedeus/LoveShare](https://github.com/breezedeus/LoveShare)。\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fbreezedeus%2Floveshare","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fbreezedeus%2Floveshare","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fbreezedeus%2Floveshare/lists"}