TrackingNet devkit

conda create -n TrackingNet python pip
pip install TrackingNet

Utility functions for TrackingNet

from TrackingNet.utils import getListSplit

# Get list of codenames for the 12 training + testing split
TrackingNetSplits = getListSplit()
print(getListSplit())
# returns ["TEST", "TRAIN_0", "TRAIN_1", "TRAIN_2", "TRAIN_3", "TRAIN_4", "TRAIN_5", "TRAIN_6", "TRAIN_7", "TRAIN_8", "TRAIN_9", "TRAIN_10", "TRAIN_11"]


# Get list of tracking sequences
print(getListSequence(split=TrackingNetSplits[1])) # return list of tracking sequences in that split
print(getListSequence(split="TEST")) # return list of tracking sequences for testing
print(getListSequence(split=["TRAIN_0", "TRAIN_1"])) # return list of tracking sequences for train splits 0 and 1
print(getListSequence(split="TRAIN")) # return list of tracking sequences for al train splits

Downloading TrackingNet

from huggingface_hub import login
login(token="<TOKEN_HUGGINGFACE>") # https://huggingface.co/settings/tokens -> Create a new token

from TrackingNet.Downloader import TrackingNetDownloader
downloader = TrackingNetDownloader(LocalDirectory="path/to/TrackingNet")

downloader.downloadSplit("TRAIN_0")
downloader.downloadSplit("TEST")

from TrackingNet.utils import getListSplit
for split in getListSplit():
    downloader.downloadSplit(split)