Closed MiTeng0215 closed 4 years ago
@MichalBusta
(e2e) mist@C-000005-GPU:~/home/E2E-MLT-master$ python demo.py -model=e2e-mlt.h5
make: Entering directory '/home/mist/home/E2E-MLT-master/nms'
make: 'adaptor.so' is up to date.
make: Leaving directory '/home/mist/home/E2E-MLT-master/nms'
Traceback (most recent call last):
File "/home/mist/home/E2E-MLT-master/net_utils.py", line 30, in load_net
v.copy_(param)
RuntimeError: The expanded size of the tensor (8400) must match the existing size (7500) at non-singleton dimension 0
Traceback (most recent call last):
File "/home/mist/home/E2E-MLT-master/net_utils.py", line 30, in load_net
v.copy_(param)
RuntimeError: The expanded size of the tensor (8400) must match the existing size (7500) at non-singleton dimension 0
e2e-mlt.h5
Using cuda ...
Awesome! nms compiled successfully! But a new problem has arisen, I don't know why the tensor has a problem.
Hi, there are 2 models provided - one with 7500 softmax and second with 8400 softmax. For demo use http://ptak.felk.cvut.cz/public_datasets/SyntText/e2e-mlt.h5 model or change model definition in code - see models.py
https://github.com/MichalBusta/E2E-MLT/blob/498e94e08912c123d0179a395dc3e4d4baba423b/models.py#L217 https://github.com/MichalBusta/E2E-MLT/blob/498e94e08912c123d0179a395dc3e4d4baba423b/models.py#L485
All the best, Michal
Now it worked. I modified the code of model.py based the #issue49.Thank you for your quick reply, thank you.
can you help me ?