CoolFace
Modelpublic

meteorain/Qwen_Qwen3-4B-Thinking-2507_int4-g128_qwen3-traces-cot-concat_2048_8_1024_256_lr0.1

sourceHugging Faceupdated 5mo agoView on Hugging Face
0likes6downloads
3 commits on main
0b630895mo ago

training logs (iter=1024 warmup=256)

meteorain
4d3220c5mo ago

DiscQuant: int4-g128 on qwen3-traces-cot-concat (w=2048 eff_bs=8 iter=1024 warmup=256)

meteorain
89a71015mo ago

initial commit

meteorain