soup
soup is Fine-tune LLMs from one YAML. Layer streaming trains an 8B model on a 4 GB laptop GPU. - MakazhanAlpamys/Soup. It is ranked #1175 on the Dev Radar, in AI dev tools, first seen 20 days ago and shared in 1 post (8.8k views).
What people said about soup on X
想动手微调个大模型，一查显存要求就凉了，仅 4 GB 的显卡按说连模型都装不进去。 偶然发现 Soup，把基座模型放在内存里，训练时一层一层轮流搬进显卡，显存里始终只装当前这一层。 靠这个办法，4 GB 显存的笔记本显卡也能微调 80 亿参数的模型，实测显存峰值 3.32 GB，训练结果和正常跑出来的一模一样。 GitHub：http://github.com/MakazhanAlpamys/Soup 用起来就一份配置文件加一条命令，批大小、模型压缩方式这些参数全自动配好，也不用远程登录什么服务器。 作者还放了个免费的 Colab 笔记本，谁都能跑一遍验证这事是真的，这点挺实诚。 拿自己的数据调个专属小模型，现在一台游戏本就够了。
— @GitHub_Daily, 20 days ago · 85 likes · see the post
Alternatives to soup
- muse.ai — New connectors are live today. Come build with us.
- classifier.dev — now outperforms jev and is free
- mimo-v2.6 RL — Streaming the run
- academy.dair.ai — Chat with Paper
- Union Alpha — Union Alpha is a multimodal model built for research, coding, and agentic workflows, while delivering frontier-level…
- cua — Draft #3943
soup in numbers
- Rank on the Dev Radar: #1175 of 1355
- Shared in 1 post by 1 account: @GitHub_Daily
- 8.8k views on those posts
- First seen 20 days ago, last shared 20 days ago
- Pricing seen by Jev: open source
- Market: AI dev tools
FAQ
What is soup?
Fine-tune LLMs from one YAML. Layer streaming trains an 8B model on a 4 GB laptop GPU. - MakazhanAlpamys/Soup It was first shared on X 20 days ago and is ranked #1175 on the Dev Radar.
Is soup free?
It is open source.
Who shared soup?
1 account on X, including @GitHub_Daily, in 1 post totalling 8.8k views.
Every post is read and classified by Jev (TypeSafe): what it is, which market it belongs to, and whether the link is a real tool. 12.2k posts from 4.7k X accounts over the last 21 days, 1.4k tools, 12 markets. Collected every 5 minutes, fully re-ranked every hour — last update 2026-09-19 18:39 UTC. Full method.