datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
VLNCE-EnvDrop
VLNCE-EnvDrop
Synthetic Vision-Language Navigation (VLN) data-augmentation set, derived from the
EnvDrop augmentation used in VLN-CE / NaVILA-style training. Each of the 146,304
samples pairs a short first-person navigation video with the natural-language
instruction the agent was following and the discrete action sequence it executed.
This dataset provides the visual + motion supervision for training a GRU-augmented
Qwen3-VL navigation model: the language conditions the… See the full description on the dataset page: https://huggingface.co/datasets/Rithvik762/VLNCE-EnvDrop.envdrop_streamvlnEnvDropTrajectoryTrajectory rendered from annotations in https://huggingface.co/datasets/cywan/StreamVLN-Trajectory-Data
NaVILA-EnvDrop