alpha parameter not present in `loss.py`

[mfournarakis]

According to the paper, the ground truth affine matrix A is normalised by the scaling constant alpha= 7.75. However, no such normalisation is present in the /src/loss.py code. Is it applied elsewhere?

[sergiomsilva]

Hi @mariosfourn , this part is present in src/sampler.py, line 13. The output matrix of that function, which is used by the loss function, already includes alpha.

[mfournarakis]

Oh I see. thanks for the clarification.

On Mon, 15 Apr 2019 at 14:36, sergiomsilva [email protected] wrote:

Hi @mariosfourn https://github.com/mariosfourn , this part is present in src/sampler.py, line 13. The output matrix of that function, which is used by the loss function, already includes alpha.

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/sergiomsilva/alpr-unconstrained/issues/66#issuecomment-483254490, or mute the thread https://github.com/notifications/unsubscribe-auth/Afj6VXv_j7YkX0GsIjKraFQFTgWK0qTJks5vhIB0gaJpZM4cv2MQ .