# Vision-language model

ViGoRL is a vision-language model fine-tuned through reinforcement learning, used to clearly associate text reasoning steps with visual coordinates to achieve precise visual reasoning and positioning.

V-JEPA 2 is a cutting-edge video understanding model developed by the FAIR team under Meta. It extends the pre-training objectives of VJEPA and has industry-leading video understanding capabilities.

Gemma is a lightweight and cutting-edge open model family launched by Google, built on the same research and technology as the Gemini model. Gemma 3 is a multimodal model that can process text and image inputs and generate text outputs.

A vision-language model specifically designed to extract structured financial information from cheque images and generate JSON-formatted output containing key information such as cheque number, payee, amount, and issue date.

A vision-language large model developed based on InternLM2 with exceptional text-image understanding and creation capabilities

xGen-MM-Vid (BLIP-3-Video) is an efficient and compact vision-language model equipped with an explicit temporal encoder, specifically designed to understand video content.

Pixtral is a multimodal model based on the Mistral architecture that can handle image and text inputs and generate text outputs.

Vision-language model initialized with OpenAI CLIP, enhanced with unsupervised adversarial fine-tuning for improved robustness

Official checkpoint of BLIP base model fine-tuned on MS-COCO dataset using MOCHA reinforcement learning framework to mitigate open-vocabulary description hallucination

MiniGPT-4 is a multimodal model that combines visual and language capabilities and is developed based on the Vicuna language model.

LLaVA is a multimodal model combining vision and language, based on the LLaMA architecture, supporting image understanding and dialogue generation.

A vision-language model based on ViT and GPT2 architectures, specifically fine-tuned for Pokémon description generation tasks

CLIP-based zero-shot and one-shot image segmentation model supporting text and image prompts

GroupViT is a vision-language model capable of performing zero-shot semantic segmentation on any given vocabulary category.