cvlab-columbia/viper
Research implementation of a visual reasoning system where an LLM generates and executes Python code to answer questions about images.
Not currently ranked — collecting fresh signals.
star history
ViperGPT combines visual understanding with language model reasoning by using an LLM to generate Python code that can process images and perform reasoning tasks. The system executes the generated code and returns results, enabling complex visual question answering. It integrates with GLIP for vision capabilities and uses OpenAI models as the code generation backbone, allowing arbitrary Python operations for visual inference.
Frequently asked
- What is cvlab-columbia/viper?
- Research implementation of a visual reasoning system where an LLM generates and executes Python code to answer questions about images.
- Is viper open source?
- Yes — cvlab-columbia/viper is an open-source project tracked on heatdrop.
- What language is viper written in?
- cvlab-columbia/viper is primarily written in Jupyter Notebook.
- How popular is viper?
- cvlab-columbia/viper has 1.7k stars on GitHub.
- Where can I find viper?
- cvlab-columbia/viper is on GitHub at https://github.com/cvlab-columbia/viper.