Datasets:

hylz
/

EvQA

Tasks:

Question Answering

Languages:

Chinese

English

ArXiv:

Tags:

event-based vision

question-answering

MCQ

Dataset card Files Files and versions

xet

Community

Dataset Viewer

The dataset viewer is not available for this subset.

Cannot get the split names for the config 'default' of the dataset.

Exception:    SplitsNotFoundError
Message:      The split names could not be parsed from the dataset config.
Traceback:    Traceback (most recent call last):
                File "/usr/local/lib/python3.12/site-packages/datasets/inspect.py", line 289, in get_dataset_config_info
                  for split_generator in builder._split_generators(
                                         ^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/packaged_modules/hdf5/hdf5.py", line 64, in _split_generators
                  with h5py.File(first_file, "r") as h5:
                       ^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/h5py/_hl/files.py", line 564, in __init__
                  fid = make_fid(name, mode, userblock_size, fapl, fcpl, swmr=swmr)
                        ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/h5py/_hl/files.py", line 238, in make_fid
                  fid = h5f.open(name, flags, fapl=fapl)
                        ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "h5py/_objects.pyx", line 56, in h5py._objects.with_phil.wrapper
                File "h5py/_objects.pyx", line 57, in h5py._objects.with_phil.wrapper
                File "h5py/h5f.pyx", line 102, in h5py.h5f.open
              FileNotFoundError: [Errno 2] Unable to synchronously open file (unable to open file: name = 'hf://datasets/hylz/EvQA@2acf4f4097bd8ae8aaf1ce5fd15e2b1637bd15bf/h5_files/EV-UAV/train_000.h5', errno = 2, error message = 'No such file or directory', flags = 0, o_flags = 0)
              
              The above exception was the direct cause of the following exception:
              
              Traceback (most recent call last):
                File "/src/services/worker/src/worker/job_runners/config/split_names.py", line 65, in compute_split_names_from_streaming_response
                  for split in get_dataset_split_names(
                               ^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/inspect.py", line 343, in get_dataset_split_names
                  info = get_dataset_config_info(
                         ^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/inspect.py", line 294, in get_dataset_config_info
                  raise SplitsNotFoundError("The split names could not be parsed from the dataset config.") from err
              datasets.inspect.SplitsNotFoundError: The split names could not be parsed from the dataset config.

Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.

Reconstruction as a Bridge for Event-Based Visual Question Answering

This is the EvQA benchmark from the paper "Reconstruction as a Bridge for Event-Based Visual Question Answering".

The corresponding code can be found at https://github.com/HYLZ-2019/EvQA.

Dataset format

The questions directory contains the annotation files in JSON format, named as questions/{dataset_name}.json. Each JSON file contains metadata for the dataset and a list of questions. The metadata includes the dataset name and description. Each entry in the question list contains the question ID, file path to the corresponding H5 file, camera type, resolution, question type, duration, keywords, and text for the questions, answers and multiple-choice options (in both English and Chinese).

The h5_files directory contains the event data files and license information, organized by dataset name. Specifically, for each dataset, the structure is h5_files/{dataset_name}/{question_id}.h5 and h5_files/{dataset_name}/LICENSE.txt.

The event data is stored in HDF5 (.h5) format, generated using h5py with compression enabled. Each H5 file contains the event stream and optional image frames.

The file attributes store essential metadata, including:

sensor_resolution: The spatial resolution (H, W) of the sensor.
num_events: The total number of events.
num_imgs: The number of intensity images (0 if unavailable).
duration: The duration of the recording in seconds.
camera_type: The model of the event camera.
data_source: The name of the source dataset.
base_time: The absolute timestamp (epoch time) of the start of the recording. If the original timestamps are relative, this is set to 0.0. The timestamps in the file are relative to this base time (i.e., starting from 0).

The events are stored separately within the H5 file:

events/xs: X-coordinates of the events (uint16).
events/ys: Y-coordinates of the events (uint16).
events/ts: Timestamps of the events in microseconds (uint64), with ts[0] == 0. We use uint64 to avoid overflow for recordings longer than approximately 1.19 hours.
events/ps: Polarity of the events (uint8), taking values of 0 or 1.

If intensity images are available, they are stored as separate datasets named images/image{idx:09d}, where idx is the image index. Each image dataset has the following attributes:

event_idx: The index of the event corresponding to the image timestamp.
timestamp: The timestamp of the image (uint64), aligned with the timestamps of the event stream.

Downloads last month: 439