chenyitian-shanshu/SIRL-Gurobi
TEXT GENERATIONConcurrency Cost:1Model Size:7.6BQuant:FP8Ctx Length:32kPublished:May 20, 2025License:mitArchitecture:Transformer0.0K Open Weights Cold

SIRL-Gurobi is a 7.6 billion parameter Qwen2.5-based language model developed by chenyitian-shanshu, specifically fine-tuned using Solver-Informed Reinforcement Learning (SIRL). This model integrates with the Gurobi optimization solver to generate accurate mathematical formulations and code for optimization problems from natural language. It excels at optimization modeling tasks, demonstrating strong performance on benchmarks like NL4OPT and MAMO.

Loading preview...