autoresearch
An automated AI research agent that enables training nano chat models on a single GPU.
- Region
- Overseas
- Pricing
- Free
- Open source
- Yes
- GitHub Stars
- ★ 85.1k
- Source
- GitHub
- Added
- 2026-06-04
- Last verified
- 2026-06-04
Overview
autoresearch is an automated AI research agent that can train nano chat models on a single GPU. It optimizes model performance by enabling AI agents to autonomously modify and iterate on training code, making it ideal for researchers aiming to improve research efficiency. Simply set up the environment and launch the agent—experiments will run automatically without human intervention.
Key features
- ▪Runs on a single GPU
- ▪Automatically modifies training code
- ▪5-minute fixed time budget
- ▪Based on val_bpb evaluation
Use cases
Pros
- ＋Improves research efficiency
- ＋Reduces manual tuning
- ＋Supports multiple AI agents
Limitations / notes
- －Requires NVIDIA GPU
- －Python 3.10+ required
Who it's for
This overview was compiled by AI from public sources and may contain inaccuracies — please refer to the official site.
FAQ
How do I get started with autoresearch?
Install the uv package manager, sync dependencies, download data, train the tokenizer, then run train.py.
Similar agents
Cursor
FeaturedAI-powered code editor with Agent mode
Dify
FeaturedOpen-source LLMOps + Agent orchestration platform
扣子 Coze
FeaturedByteDance's no-code Agent/Bot building platform
Trae
FeaturedAI IDE (with Agent) from ByteDance
LangChain
FeaturedThe most mainstream LLM/Agent orchestration framework
Devin
FeaturedAutonomous software engineer Agent by Cognition
Something wrong? Let us know on the About page and we'll fix it.