TIGER-Lab/VisCoder2-7B

VisCoder2-7B by TIGER-Lab is a lightweight multi-language visualization coding model based on Qwen2.5-Coder-7B-Instruct, designed for executable code generation, rendering, and iterative self-debugging. It is trained on the VisCode-Multi-679K dataset, covering 12 programming languages for visualization tasks. This model excels at generating semantically consistent visual outputs by aligning natural language instructions with rendering results. Its primary strength lies in multi-language visualization code generation with self-debugging capabilities.