devtaji
/

OpenThinker-Agent-repro-RL

Model card Files Files and versions

OpenThinker-Agent-repro-RL / baseline-lr1e-6-step20 /tokenizer.json

Commit History

Preserve lr=1e-6 RL baseline (step 20) before lr=1e-5 resume

1cd6da2
verified

devtaji commited on Jun 27