https://github.com/OpenBMB/MiniCPM

MiniCPM3-4B: An edge-side LLM that surpasses GPT-3.5-Turbo.
https://github.com/OpenBMB/MiniCPM

Last synced: about 1 month ago
JSON representation

MiniCPM3-4B: An edge-side LLM that surpasses GPT-3.5-Turbo.

• Host: GitHub
• URL: https://github.com/OpenBMB/MiniCPM
• Owner: OpenBMB
• License: apache-2.0
• Created: 2024-01-29T08:21:15.000Z (11 months ago)
• Default Branch: main
• Last Pushed: 2024-09-07T07:53:17.000Z (3 months ago)
• Last Synced: 2024-09-09T03:35:35.592Z (3 months ago)
• Language: Python
• Homepage:
• Size: 143 MB
• Stars: 5,590
• Watchers: 63
• Forks: 369
• Open Issues: 26
• Metadata Files:
  • Readme: README-en.md
  • License: LICENSE

Awesome Lists containing this project

• ai-game-devtools - MiniCPM-2B - side LLM outperforms Llama2-13B. | | | Tool | (Project List / <span id="tool">Tool (AI LLM)</span>)
• awesome-multi-modal - https://github.com/OpenBMB/MiniCPM
• Awesome-Text2SQL - [code
• awesome-multi-modal - https://github.com/OpenBMB/MiniCPM
• AiTreasureBox - OpenBMB/MiniCPM - 12-07_7164_1](https://img.shields.io/github/stars/OpenBMB/MiniCPM.svg)|MiniCPM-2B: An end-side LLM outperforms Llama2-13B.| (Repos)
• StarryDivineSky - OpenBMB/MiniCPM - 2B 仅有 24亿的非词嵌入参数量, 总计2.7B参数量。经过 SFT 后，在公开综合性评测集上，与 Mistral-7B相近（中文、数学、代码能力更优），整体性能超越 Llama2-13B、MPT-30B、Falcon-40B 等模型。经过 DPO 后，在当前最接近用户体感的评测集 MTBench上，也超越了 Llama2-70B-Chat、Vicuna-33B、Mistral-7B-Instruct-v0.1、Zephyr-7B-alpha 等众多代表性开源大模型。以 MiniCPM-2B 为基础构建端侧多模态大模型 MiniCPM-V，整体性能在同规模模型中实现最佳，超越基于 Phi-2 构建的现有多模态大模型，在部分评测集上达到与 9.6B Qwen-VL-Chat 相当甚至更好的性能。经过 Int4 量化后，可在手机上进行部署推理，流式输出速度略高于人类说话速度。也直接跑通了多模态大模型在手机上的部署。一张1080/2080可高效参数微调，一张3090/4090可全参数微调，一台机器可持续训练 MiniCPM，二次开发成本较低。 (A01_文本生成_文本对话 / 大语言对话模型及数据)
• awesome-ai-papers - [MiniCPM - MoE](https://github.com/SkyworkAI/Skywork-MoE)\]\[[Orion](https://github.com/OrionStarAI/Orion)\]\[[BELLE](https://github.com/LianjiaTech/BELLE)\]\[[Yuan-2.0](https://github.com/IEIT-Yuan/Yuan-2.0)\]\[[Yuan2.0-M32](https://github.com/IEIT-Yuan/Yuan2.0-M32)\]\[[Fengshenbang-LM](https://github.com/IDEA-CCNL/Fengshenbang-LM)\]\[[Index-1.9B](https://github.com/bilibili/Index-1.9B)\]\[[Aquila2](https://github.com/FlagAI-Open/Aquila2)\] (NLP / 3. Pretraining)
• awesome-ai-papers - [MiniCPM - MoE](https://github.com/SkyworkAI/Skywork-MoE)\]\[[Orion](https://github.com/OrionStarAI/Orion)\]\[[BELLE](https://github.com/LianjiaTech/BELLE)\]\[[Yuan-2.0](https://github.com/IEIT-Yuan/Yuan-2.0)\]\[[Yuan2.0-M32](https://github.com/IEIT-Yuan/Yuan2.0-M32)\]\[[Fengshenbang-LM](https://github.com/IDEA-CCNL/Fengshenbang-LM)\]\[[Index-1.9B](https://github.com/bilibili/Index-1.9B)\]\[[Aquila2](https://github.com/FlagAI-Open/Aquila2)\] (NLP / 3. Pretraining)