test

Public synthetic test dataset for the epinet training/data-loading code.

This repository contains generated random .npy samples only. It does not contain real clinical, patient, or research data.

Contents

• synthetic_epinet_test.csv: sample index used by train.py
• data/synthetic_epinet_0000.npy to data/synthetic_epinet_0449.npy
• Each sample is a NumPy float32 array with shape (29, 2000)
• Total data payload is about 100 MiB

Minimal usage

From an epinet checkout, point the code at this clone:

EPINET_DATA_FILE=/path/to/test/synthetic_epinet_test.csv \
EPINET_DATA_ROOT=/path/to/test \
EPINET_N_FOLD=1 \
EPINET_EPOCHS=1 \
EPINET_BATCH_SIZE=8 \
EPINET_VALID_BATCH_SIZE=8 \
EPINET_PROCESS_NUM=0 \
python train.py

This dataset is intended for smoke tests and pipeline checks, not model quality.