PyPI page
Author:
rlcli contributors
Summary:
One-shot CLI for Tinker-API post-training on your own GPUs (SkyRL backend): serve, train, and fused RL losses like forward_backward('gspo') in a single command.
Latest version:
0.1.0
Required dependencies:
click
|
httpx
|
tinker
Optional dependencies:
math-verify
|
modal
|
pylatexenc
|
pytest
|
sympy
Downloads last day:
1
Downloads last week:
10
Downloads last month:
29
7-day smoothing