The rank-8, alpha-16 SFT LoRA adapter was merged into the base model with PEFT merge_and_unload, then saved as six safetensors shards. This merged model is intended to serve as a standalone SFT model and as the parent model for subsequent GRPO experiments.

kyleliu789/qwen3-14b-svamp14-sft-merged-bf16-r8-a16

Qwen3-14B SVAMP14 SFT — Merged BF16 Rank-8/A16

kyleliu789/
qwen3-14b-svamp14-sft-merged-bf16-r8-a16