Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/multimodal-art-projection/MAP-NEO
https://github.com/multimodal-art-projection/MAP-NEO
Last synced: about 1 month ago
JSON representation
- Host: GitHub
- URL: https://github.com/multimodal-art-projection/MAP-NEO
- Owner: multimodal-art-projection
- Created: 2024-05-08T11:58:35.000Z (8 months ago)
- Default Branch: main
- Last Pushed: 2024-06-21T02:36:44.000Z (6 months ago)
- Last Synced: 2024-08-03T09:06:55.487Z (5 months ago)
- Language: Python
- Size: 103 MB
- Stars: 797
- Watchers: 10
- Forks: 73
- Open Issues: 25
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
- StarryDivineSky - multimodal-art-projection/MAP-NEO - NEO 是一个完全开源的大型语言模型,包括预训练数据、数据处理管道 (Matrix)、预训练脚本和对齐代码。它在 4.5T 中英文词元上从头开始训练,表现出与 LLaMA2 7B 相当的性能。MAP-Neo 模型在推理、数学和编码等具有挑战性的任务中提供类似专有模型的性能,优于同等规模的同类产品。出于研究目的,我们的目标是在LLM培训过程中实现完全透明。为此,我们全面发布了 MAP-Neo,包括最终和中间检查点、自训练标记器、预训练语料库,以及高效、稳定优化的预训练代码库。 (A01_文本生成_文本对话 / 大语言对话模型及数据)