Forums

Image Restoration for Under-display Camera @MIPI-challenge Forum

Go back to competition Back to thread list Post in this thread

> Submission Failed - RuntimeError: expected scalar type Double but found Float

RuntimeError: expected scalar type Double but found Float.
The result is stored as int16 numpy files.
When checked in the evaluation script after commands below is of dtype torch.float64/torch.Double
res_img = tone_map(np.load(res_img))
ref_img = tone_map(np.load(ref_img))
res_img = torch.tensor(res_img).permute(2, 0, 1).unsqueeze_(0)
ref_img = torch.tensor(ref_img).permute(2, 0, 1).unsqueeze_(0).

Posted by: hrishikeshps94 @ June 16, 2022, 7:22 a.m.

Hi @hrishikeshps94,

You should store the results as Float type rather than int16 or uint8. Please refer to https://codalab.lisn.upsaclay.fr/forums/4874/653/.

Best,
Organizers

Posted by: jnjaby @ June 16, 2022, 8 a.m.
Post in this thread