AgriJev Family of Qwen3.5 Models for Agricultural Candidate Scoring
October 15, 2026
AgriJev adapts Qwen3.5 models into sizes ranging from 0.8B to 9B to perform multi-scale agricultural candidate scoring in a single forward pass. The 9B model achieves 75.44% mean accuracy across 17 test views, outperforming original and standard SFT generators.
HOW THIS AFFECTS YOU
●
builderYou can deploy specialized, small-scale models for high-accuracy agricultural decision-making tasks.
●
researcherThis demonstrates the effectiveness of multi-scale adapted models for heterogeneous entity scoring.