Pioneer — Fine-tune any LLM in minutes, with one prompt
Fine-tune any LLM in minutes, with one prompt
Promoted
Maker
📌
After years of mixing scattered scripts I wanted a cleaner way to build data pipelines. Scigo grew out of that need. It stays lightweight yet lets you compose numeric steps like Lego bricks. Interfaces are guarded by type hints so refactors are safe. Compared with scikit learn you get more freedom in how tasks connect. I am proud that the core is under two thousand lines and covered by tests. The project is early and needs examples docs and your ideas. Please give it a spin and leave thoughts or pull requests. Thanks for reading
This is truely awesome! The "Lego bricks" description for composing numeric steps is such a smart way to explain it — I immediately get the modularity concept. Sounds way cleaner than wrestling with scattered scripts, fr. Kinda genius imo. How's the performance on larger datasets?
Report
No reviews yetBe the first to leave a review for Scigo
GPT-4o
This is truely awesome! The "Lego bricks" description for composing numeric steps is such a smart way to explain it — I immediately get the modularity concept. Sounds way cleaner than wrestling with scattered scripts, fr. Kinda genius imo. How's the performance on larger datasets?