InternVL2-8B-MPO
InternVL2-8B-MPO enhances multimodal reasoning and accuracy, reducing hallucinations for better task performance and deployment flexibility.
What is InternVL2-8B-MPO
InternVL2-8B-MPO is a powerful multimodal large language model enhanced with Mixed Preference Optimization for improved multimodal reasoning capabilities. It achieves high accuracy on MathVista and reduces hallucinations compared to its predecessor. Ideal for researchers and developers needing to analyze multimodal data and improve inference tasks.