Modelpublic

RL-Forgetting-Experiments-3/qwen2.5-3b-kk-sft-replay-uniform-lam1-step2222

sourceHugging Faceapache-2.0updated 2d agoView on Hugging Face

0likes18downloads

Model Card

Qwen2.5-3B KK replay-SFT: uniform, step 2222

Status: Intermediate checkpoint. The restored fork continued past this point; this is the latest fully inventoried valid model, not a completed step-3175 sweep.

Source result: 7209866942.90703-14e439f4-2041 Source run: qwen3b_kk_ordered_lr1e-5_bs256_seed42_rpuniform_lam1.0_sync1_h100_fork635_20260922