# Zero-shot Generalization

VisualCloze is a universal image generation framework based on visual context learning, supporting generalization across multiple in-domain tasks and unseen tasks, generating target images and intermediate results in a single step.

PoseLess is an innovative robotic hand control framework that directly maps 2D images to joint angles using projection representations, eliminating the need for explicit pose estimation.

Poseless-3B is a vision-language model (VLM)-based robotic hand control framework that directly maps 2D images to joint angles without explicit pose estimation.

A visual retrieval model based on Qwen2.5-VL-3B-Instruct and ColBERT strategy, capable of generating multi-vector representations for text and images to enable efficient document retrieval.

A visual retrieval model based on Qwen2-VL-2B-Instruct and ColBERT strategy, capable of efficiently indexing documents through visual features

A foundational model for promptable visual segmentation in images and videos developed by FAIR

OpenVLA 7B is an open-source vision-language-action model trained on the Open X-Embodiment dataset, capable of generating robot actions based on language instructions and camera images.

OpenVLA v0.1 7B is an open-source vision-language-action model trained on the Open X-Embodiment dataset, supporting various robot controls.

KRISSBERT is a knowledge-enhanced self-supervised learning model for biomedical entity linking. It trains contextual encoders using unannotated text and domain knowledge to effectively address the diversity and ambiguity of entity names.

A fine-tuned text-to-SQL conversion model based on T5-3B architecture, significantly improving structured query generation accuracy through PICARD constrained decoding