datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
talent_plus_rl_groups_of_50_with_audiobox_scoresnew-rl-vitarlbenchBridgeVLA_RLBench_EVAL_DATAarxiv: https://arxiv.org/abs/2506.07961
BridgeVLA_RLBench_TRAIN_DATAarxiv: https://arxiv.org/abs/2506.07961
BridgeVLA_RLBench_TRAIN_BUFFERarxiv: https://arxiv.org/abs/2506.07961
found-rl_dataset
Found-RL Dataset: Demonstration Data for VLM Fine-tuning
🚗 Dataset Overview
This dataset contains large-scale demonstration data collected from the CARLA simulator, designed to fine-tune Vision-Language Models (VLMs) for autonomous driving tasks. It serves as the data foundation for the paper "Found-RL: Foundation Model-Enhanced Reinforcement Learning for Autonomous Driving".
The dataset comprises approximately 1.374 million state-action transitions collected across… See the full description on the dataset page: https://huggingface.co/datasets/ys-qu/found-rl_dataset.sd_ocr_rl_embeddingsmininet-1krlbench_taskrlbench
RLBench Dataset
RLBench 18 tasks following the PerAct protocol, with image resolution 128×128.
100 demonstrations per task for training.
25 demonstrations per task for testing.
video-rlhamster_rlbenchVideo-RLflux_rl_embeddingssd_geneval_rl_embeddingsRLVE_env400_d0-15MPC_RLcollect_vlm_v3_1_rl-0521-backupPyVision-RL-Eval-Samples