CoolFace
Modelpublic

Brain2nd/NeuronSpark-V3-1.1B-SFT-V25-NoThink

sourceHugging Faceapache-2.0updated 5mo agoView on Hugging Face
0likes18downloads
Model Card

NeuronSpark V3 — 1.1B SFT V2.5 NoThink (step 2000)

非 think SFT (V2.5 数据集, 292,525 samples, 0 think tags). bs=12 × maxlen=1024 × 8 GPU effbatch=768. Base = pretrain step 108000.