agurung/lcb-med-hard-ft-qwen3-4b-dft-mixed-24-highlr-v1

agurung/lcb-med-hard-ft-qwen3-4b-dft-mixed-24-highlr-v1

PEFT LoRA adapter archive for the LCB med/hard Qwen3-4B FT experiment.

• —Base model: Qwen/Qwen3-4B-Instruct-2507
• —Local run name: dft_mixed_24-lora-r128-a32-lr5e-4-const-lr5e-4-qps8-gpuauto-ep4-highlr_v1
• —Arm: dft_mixed_24
• —Variant: highlr-v1
• —Adapter checkpoints: 16
• —Included bytes: 18164814939
• —Excluded local optimizer/RNG bytes: 0

The archive intentionally keeps HF/PEFT model artifacts and training metadata, but excludes optimizer, scheduler, and RNG state files. These repos are meant for evaluation and RL seeding, not exact optimizer-state resume.

Checkpoint dirs:

checkpoint-114, checkpoint-133, checkpoint-152, checkpoint-171, checkpoint-19, checkpoint-190, checkpoint-209, checkpoint-228, checkpoint-247, checkpoint-266, checkpoint-285, checkpoint-300, checkpoint-38, checkpoint-57, checkpoint-76, checkpoint-95