why my loss always keep in 300 and acc under in 50%?

[ZengyuanYu]

苏AZ3A25 1.0 <-> 苏AZ3A25 1.0 苏AJ3H16 1.0 <-> 苏AJ3H16 1.0 苏AC0378 1.0 <-> 苏AS0377 1.0 苏AY9L95 1.0 <-> 苏AX9L15 1.0 苏AZ1W57 1.0 <-> 苏AZ7X17 1.0 苏AR3595 1.0 <-> 苏AR0195 1.0 苏AI0P50 1.0 <-> 苏AI0P59 1.0 苏AE8J34 1.0 <-> 苏AE8J34 1.0 苏AO9R64 1.0 <-> 苏AO9R64 1.0 苏AC5N78 1.0 <-> 苏AC5N78 1.0 苏AU4A12 1.0 <-> 苏AU4A12 1.0 苏AN4780 1.0 <-> 苏AN4780 1.0 B77120 42.00% 100.00% loss: 552.0714111328125 (digits: 552.0714111328125, presence: 1.8921246439497583e-15) |XX X X XXXX XX XXXX XX XXX XX XXX XXXXX | time for 60 batches 9.751856088638306

I had train it 3 hours

[wenting521]

Can you share your code?Thank you very much!

[KassemKallas]

Hi Sir,

I have the same problem! did you solve it?

[wendaomumao]

我在第三步，识别的时候遇到了这样的问题 deep-anpr/detect.py:102: RuntimeWarning: invalid value encountered in greater for window_coords in numpy.argwhere(y_val[0, :, :, 0] > -math.log(1./0.99 - 1)): 你遇到了没有曾大佬

[Utopiaaaaaaa]

did you solve the problem? I have a similar problem. I follow the tutorial completely, training 7 hours, the result is that.