Robotic Manipulation & Task Learning
23 repos
Systems and benchmarks for training robotic agents to perform manipulation tasks through learning-based approaches. The cluster centers on LIBERO, a suite of standardized benchmark environments for evaluating robot learning across object, spatial, and goal-based manipulation tasks, with both standard and professional variants. These repos emphasize reproducible evaluation of vision-language models and reinforcement learning methods for robotics, using efficient tensor serialization (safetensors) and modular code structures to support research in generalizable robot control.
safetensors ·360
robotics ·352
custom_code ·334
openvla ·334
transformers ·297
pretraining ·288
image-text-to-text ·288
multimodal ·288
feature-extraction ·279
Vision-Language-Action ·55
