This release packages a benchmark-ready widefield calcium imaging dataset for neural time-series forecasting and Nethobench-style evaluation.
The release contains two complementary data representations:
A prepared benchmark tensor: data/data100_ba16.npy, already organized into fixed-length subsequences for model training and evaluation.
An unprepared source Parquet table: data/data-clean-all.parquet, containing the… See the full description on the dataset page: https://huggingface.co/datasets/Nethobench/nethobench-widefield-v1.
