Я пытаюсь изменить внутреннюю архитектуру сети EAST с ResNet-50 на MobileNet.Я смог обучить модель, но когда я попытался протестировать ее, я получил следующее сообщение об ошибке:
File "eval.py", line 193, in <module>
main()
File "eval.py", line 154, in main
score_map, geo_map = model.predict(img_resized[np.newaxis, :, :, :])
File "/home/niddal/.local/lib/python3.6/site-packages/keras/engine/training.py", line 1169, in predict
steps=steps)
File "/home/niddal/.local/lib/python3.6/site-packages/keras/engine/training_arrays.py", line 294, in predict_loop
batch_outs = f(ins_batch)
File "/home/niddal/.local/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py", line 2715, in __call__
return self._call(inputs)
File "/home/niddal/.local/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py", line 2675, in _call
fetched = self._callable_fn(*array_vals)
File "/usr/lib/python3/dist-packages/tensorflow/python/client/session.py", line 1439, in __call__
run_metadata_ptr)
File "/usr/lib/python3/dist-packages/tensorflow/python/framework/errors_impl.py", line 528, in __exit__
c_api.TF_GetCode(self.status.status))
tensorflow.python.framework.errors_impl.InvalidArgumentError: Input to reshape is a tensor with 552960 values, but the requested shape has 393216
[[{{node reshape_1/Reshape}}]]
Есть предложения о том, как изменить вход?