pickfit hasn’t verified Retrograd yet, so there are no tested install steps.
What follows comes from its README only.
Use it
Before you install: known vulnerabilities
pickfit found 1 known high or critical vulnerability in the packages a fresh install brings (OSV.dev, 3 Nov 2026). Weigh this before you install it.
crates:pyo3 0.25.1 PyO3 has an Out-of-bounds Read in `nth` / `nth_back` for `PyList` and `PyTuple` iterators (GHSA-36hh-v3qg-5jq4, high, fixed in 0.29.0)
You need first
A GPU is optional
What it does
Good for
train a LoRA adapter for a GGUF base model
fine-tune a model with SFT on text or chat
run preference optimization with DPO
run multi-turn agentic GRPO with MCP tools
serve a trained adapter via OpenAI-compatible API
Quality score
Out of 100: how well its install and main features worked when pickfit tried them, how complete its install docs are, and how well it is maintained.
Out of 100, from public counts against other tools of the same kind: GitHub stars and forks, release and package downloads, Hacker News points. Low is not bad: a good tool can still be little known, and pickfit tests it the same way.
Feature checks
pickfit tries the main features the README advertises, one by one, in the sandbox. Some need an account, a GPU or a display the sandbox doesn’t have, so they stay unchecked, and the card says which and why.
Hidden gem
A verified tool that is well made and not yet widely known.