{"full_name":"hpcaitech/ColossalAI","owner":"hpcaitech","name":"ColossalAI","url":"https://github.com/hpcaitech/ColossalAI","description":"Making large AI models cheaper, faster and more accessible","homepage":"https://www.colossalai.org","topics":[],"language":"Python","license":"Apache-2.0","avatar":"https://avatars.githubusercontent.com/u/88699314?v=4","created_at":"2021-10-28T16:19:44Z","pushed_at":"2026-11-10T08:25:13Z","metrics":{"stars":41431,"forks":4483,"watchers":374,"issues":1771,"prs":4472,"releases":50,"issues_open":null,"issues_closed":null,"prs_open":null,"prs_closed":null,"prs_merged":null,"discussions":215},"growth":{"day":{"stars":0,"forks":-2,"base_day":"2026-11-09"}},"history":{"dates":["2026-11-09","2026-11-10"],"stars":[41431,41431],"forks":[4485,4483],"watchers":[374,374],"issues_open":[null,null],"prs_open":[null,null]},"boards":{},"same_owner":[{"full_name":"hpcaitech/Open-Sora","name":"Open-Sora","boards":{}}],"issues":[{"number":3331,"title":"Can not train llama-7b due to OOM on 40GA100","comments":48,"created_at":"2023-03-30T03:03:45Z"},{"number":4008,"title":"代码bugs太多了[BUG]: ","comments":44,"created_at":"2023-06-16T02:25:38Z"},{"number":4578,"title":"[BUG]: 使用llama2增量预训练失败","comments":39,"created_at":"2023-09-01T02:00:45Z"},{"number":3565,"title":"[FEATURE]: ChatGLM model support","comments":24,"created_at":"2023-04-14T10:27:11Z"},{"number":6066,"title":"[DOC]: 环境安装失败","comments":19,"created_at":"2024-09-21T18:23:35Z"},{"number":2309,"title":"[FEATURE]: Any plan to support train_dreambooth_colossalai with train_text_encoder?","comments":18,"created_at":"2023-01-04T06:12:07Z"},{"number":6110,"title":"[BUG]: Llama3.1 HybridParallelPlugin train failed when pp_size>1","comments":17,"created_at":"2024-11-02T08:00:36Z"},{"number":3440,"title":"使用2台服务器共10个GPU训练，反而比使用1个服务器5个GPU训练耗时长","comments":16,"created_at":"2023-04-04T08:49:33Z"}],"discussions":[{"number":6350,"title":"RuntimeError: CUDA driver error: invalid argument","comments":1,"category":"Community | Q&A","created_at":"2025-06-17T09:10:26Z"},{"number":6342,"title":"Environment","comments":1,"category":"Community | General","created_at":"2025-06-09T02:30:04Z"},{"number":6244,"title":"lora_finetune.py for continuous pretraining for DeepSeek-V3","comments":2,"category":"Community | Q&A","created_at":"2025-03-11T12:51:33Z"},{"number":6243,"title":"AI Agent 中文技术交流群","comments":6,"category":"Community | General","created_at":"2025-03-10T14:21:21Z"},{"number":6226,"title":"【R1 SFT Bug，loss should start from 1】","comments":0,"category":"Community | Show and tell","created_at":"2025-02-27T08:15:20Z"},{"number":6194,"title":"Using HybridParallelPlugin can I do [2D, 2.5D, 3D] tensor parallelism","comments":2,"category":"Community | General","created_at":"2025-02-17T11:37:00Z"},{"number":6190,"title":"Does ColossalAI support Huawei Ascend hardware platform for trainning?","comments":1,"category":"Community | Q&A","created_at":"2025-02-13T09:51:24Z"},{"number":6182,"title":"文档写是垃圾","comments":1,"category":"Community | General","created_at":"2025-01-20T17:22:39Z"}],"latest_release":{"name":"Version v0.5.0 Release Today!","tag":"v0.5.0","published_at":"2025-06-04T06:00:47Z","body":"## What's Changed\r\n* [HotFix] update load lora model Readme; by @duanjunwen in https://github.com/hpcaitech/ColossalAI/pull/6240\r\n* Update README.md by @Yanjia0 in https://github.com/hpcaitech/ColossalAI/pull/6268\r\n* [ci] update ci by @flybird11111 in https://github.com/hpcaitech/ColossalAI/pull/6254\r\n* [upgrade]Upgrade transformers by @flybird11111 in https://github.com/hpcaitech/ColossalAI/pull/6320\r\n* [release] release version by @flybird11111 in https://github.com/hpcaitech/ColossalAI/pull/6330\r\n\r\n\r\n**Full Changelog**: https://github.com/hpcaitech/ColossalAI/compare/v0.4.9...v0.5.0"},"readme":true,"readme_zh":false,"news":[],"stale":false,"snapshot_at":"2026-11-10T09:35:25Z","history_start":"2026-09-06"}