Is it possible to run on CPU?

[orilifs]

Hello Amazing work @chenxin-dlut :) much appreciated.

I am interested to run TransT on CPU only machine, if possible.

I am wondering if changing the parameters here:

from pytracking.utils import TrackerParams
from pytracking.features.net_wrappers import NetWithBackbone

def parameters():
    params = TrackerParams()
    params.debug = 0
    params.visualization = False
    params.use_gpu = True
    params.net = NetWithBackbone(net_path='transt.pth',
                                 use_gpu=params.use_gpu)
    return params

params.use_gpu = False Will this even run? (i.e., incase it runs something like IoUNet for bbox regression that currently only implemented on cuda-gpu) Will other params or code need to be changed please?

Thank you for the time and answers :)

[chenxin-dlut]

There are no IoUNet–like component in this method, I think it should not be complex to runned on CPU