Epoch 1/10
42875/42875 [==============================] - 822s 19ms/step - loss: 2.1951 - sparse_categorical_accuracy: 0.6136
Epoch 2/10
42875/42875 [==============================] - 804s 19ms/step - loss: 0.6784 - sparse_categorical_accuracy: 0.8422
Epoch 3/10
42875/42875 [==============================] - 804s 19ms/step - loss: 0.4639 - sparse_categorical_accuracy: 0.8804
Epoch 4/10
42875/42875 [==============================] - 805s 19ms/step - loss: 0.3498 - sparse_categorical_accuracy: 0.9040
Epoch 5/10
42875/42875 [==============================] - 803s 19ms/step - loss: 0.2857 - sparse_categorical_accuracy: 0.9173
Epoch 6/10
42875/42875 [==============================] - 805s 19ms/step - loss: 0.2349 - sparse_categorical_accuracy: 0.9322
Epoch 7/10
42875/42875 [==============================] - 804s 19ms/step - loss: 0.2088 - sparse_categorical_accuracy: 0.9384
Epoch 8/10
42875/42875 [==============================] - 804s 19ms/step - loss: 0.1866 - sparse_categorical_accuracy: 0.9452
Epoch 9/10
42875/42875 [==============================] - 805s 19ms/step - loss: 0.1799 - sparse_categorical_accuracy: 0.9469
Epoch 10/10
42875/42875 [==============================] - 805s 19ms/step - loss: 0.1563 - sparse_categorical_accuracy: 0.9536

>>> print(np.sum(test_y == predicts) / test_y.shape[0])
0.89

>>> print(F1 = f1_m(predicts,test_y))
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "<stdin>", line 2, in f1_m
  File "<stdin>", line 2, in precision_m
  File "C:\Python36\lib\site-packages\keras\backend\tensorflow_backend.py", line 1601, in clip
    min_value = _to_tensor(min_value, x.dtype.base_dtype)
AttributeError: 'numpy.dtype' object has no attribute 'base_dtype'
>>>
>>>