CoolFace
Modelpublic

RL-Forgetting-Experiments-3/qwen2.5-3b-kk-sft-replay-ce-lam1-step952

sourceHugging Faceapache-2.0updated 3d agoView on Hugging Face
0likes13downloads
Model Card

Qwen2.5-3B KK replay-SFT: ce, step 952

Status: Intermediate checkpoint. The last verified canonical CE training-metrics file was incomplete; this repository makes no step-3175 completion claim.

Source result: 7209966162.62573-2caa9529-2651 Source run: qwen3b_kk_ordered_lr1e-5_bs256_seed42_rpce_lam1.0_gcr_20260921