git
ask
hub
Privacy
Terms
Sign in with GitHub
vivekvkashyap/synthetic-self-improve-rl
↗
Stars ·
26
License ·
MIT
Simplify and Visualize This Repo
Scan the safety of this repo
Self Host this repo
Check out more work of Developer: vivekvkashyap
Ask anything about this repo to start.
Run the synthetic-self-improve-rl skill on gsm8k with Qwen3-0.6B and a 4 hour budget
Adapt the skill to point at a custom Prime hub environment for code generation
Add a flag to control the number of synthetic rows generated per iteration
Replace prime-rl with TRL as the training backend and keep the same loop
Plot the eval accuracy curve across iterations from the skill output logs
Full explanation on explaingit →
📎
Send
By chatting or signing in you agree to the
Terms
and chat-message logging (revocable in
History
).