AIPlans/qwen3-8b-ipo-hh-rlhf
Text Generation
•
Updated
•
4
Post Training Versions - Qwen 0.6B
Different versions of Qwen 0.6b, where the only difference is the post training method used. The post training database should be the hh rlhf dataset.
-
-
AIPlans/qwen3-0.6b-dpo-lora
Text Generation • 0.6B • Updated • 4 • 1
AIPlans/qwen3-0.6b-hh-rlhf-sft
0.6B • Updated • 19Note Full SFT on HH-RLHF dataset (Helpful, Harmless, Honest). All 600M parameters fine-tuned for 3 epochs.