我正在尝试双向模型：
model = Sequential()
model.add(Bidirectional(LSTM(256, input_shape=(X_train_modified.shape[1], X_train_modified.shape[2]), return_sequences=True)))
model.add(Bidirectional(LSTM(256)))
model.add(Dense(ys_train.shape[1], activation='softmax'))
sgd = tf.keras.optimizers.SGD(lr=0.01, decay=1e-6, momentum=0.9, nesterov=True)
model.compile(loss='categorical_crossentropy',
optimizer=sgd,
metrics=['accuracy'])
我使用ReduceLROnPlateau： reduce_lr = tf.keras.callbacks.ReduceLROnPlateau（monitor ='val_loss'，factor = 0.2， 耐心= 5，min_lr = 0.001）
if a metric stopped improving
try:
history = model.fit(X_train_modified, ys_train, epochs = 500, verbose=2,batch_size=16000,validation_data=(X_test_modified, ys_test),
callbacks=[reduce_lr] )
from keras.models import load_model
file_Name= "/content/drive/My Drive/Colab Notebooks/saved_models/ganjoor_embdding64_bidirectional_LSTM400_softmax.h5"
model.save(file_Name)
except:
model.save(file_Name)
但是，结果并没有超过设置为5个时期的患者。
Train on 969610 samples, validate on 646408 samples
Epoch 1/500
969610/969610 - 141s - loss: 3.4363 - acc: 0.1211 - val_loss: 2.9694 - val_acc: 0.1235
Epoch 2/500
969610/969610 - 135s - loss: 2.8853 - acc: 0.1232 - val_loss: 2.8558 - val_acc: 0.1235
Epoch 3/500
969610/969610 - 135s - loss: 2.8474 - acc: 0.1232 - val_loss: 2.8446 - val_acc: 0.1235
Epoch 4/500
969610/969610 - 135s - loss: 2.8414 - acc: 0.1232 - val_loss: 2.8414 - val_acc: 0.1235
Epoch 5/500
969610/969610 - 135s - loss: 2.8392 - acc: 0.1232 - val_loss: 2.8399 - val_acc: 0.1235
Epoch 6/500
969610/969610 - 135s - loss: 2.8380 - acc: 0.1232 - val_loss: 2.8389 - val_acc: 0.1235
Epoch 7/500
969610/969610 - 135s - loss: 2.8372 - acc: 0.1232 - val_loss: 2.8384 - val_acc: 0.1235
Epoch 8/500
969610/969610 - 135s - loss: 2.8367 - acc: 0.1232 - val_loss: 2.8379 - val_acc: 0.1235
Epoch 9/500
969610/969610 - 135s - loss: 2.8363 - acc: 0.1232 - val_loss: 2.8375 - val_acc: 0.1235
Epoch 10/500
969610/969610 - 135s - loss: 2.8360 - acc: 0.1232 - val_loss: 2.8372 - val_acc: 0.1235
Epoch 11/500
969610/969610 - 135s - loss: 2.8358 - acc: 0.1232 - val_loss: 2.8371 - val_acc: 0.1235
Epoch 12/500
969610/969610 - 135s - loss: 2.8356 - acc: 0.1232 - val_loss: 2.8369 - val_acc: 0.1235
Epoch 13/500
969610/969610 - 135s - loss: 2.8355 - acc: 0.1232 - val_loss: 2.8368 - val_acc: 0.1235
Epoch 14/500
969610/969610 - 136s - loss: 2.8353 - acc: 0.1232 - val_loss: 2.8367 - val_acc: 0.1235
Epoch 15/500
969610/969610 - 137s - loss: 2.8352 - acc: 0.1232 - val_loss: 2.8365 - val_acc: 0.1235
Epoch 16/500
969610/969610 - 139s - loss: 2.8351 - acc: 0.1232 - val_loss: 2.8364 - val_acc: 0.1235
Epoch 17/500
969610/969610 - 139s - loss: 2.8350 - acc: 0.1232 - val_loss: 2.8363 - val_acc: 0.1235
Epoch 18/500
969610/969610 - 139s - loss: 2.8349 - acc: 0.1232 - val_loss: 2.8363 - val_acc: 0.1235
Epoch 19/500
969610/969610 - 139s - loss: 2.8349 - acc: 0.1232 - val_loss: 2.8362 - val_acc: 0.1235
Epoch 20/500
969610/969610 - 139s - loss: 2.8348 - acc: 0.1232 - val_loss: 2.8361 - val_acc: 0.1235
Epoch 21/500
969610/969610 - 138s - loss: 2.8347 - acc: 0.1232 - val_loss: 2.8361 - val_acc: 0.1235
Epoch 22/500
969610/969610 - 138s - loss: 2.8347 - acc: 0.1232 - val_loss: 2.8361 - val_acc: 0.1235
Epoch 23/500
969610/969610 - 135s - loss: 2.8346 - acc: 0.1232 - val_loss: 2.8360 - val_acc: 0.1235
Epoch 24/500
...
Epoch 44/500
969610/969610 - 136s - loss: 2.8339 - acc: 0.1232 - val_loss: 2.8353 - val_acc: 0.1235
Epoch 45/500
969610/969610 - 136s - loss: 2.8339 - acc: 0.1232 - val_loss: 2.8353 - val_acc: 0.1235
Epoch 46/500
969610/969610 - 135s - loss: 2.8339 - acc: 0.1232 - val_loss: 2.8353 - val_acc: 0.1235
Epoch 47/500
969610/969610 - 137s - loss: 2.8339 - acc: 0.1232 - val_loss: 2.8353 - val_acc: 0.1235
Epoch 48/500
969610/969610 - 138s - loss: 2.8339 - acc: 0.1232 - val_loss: 2.8353 - val_acc: 0.1235
Epoch 49/500
969610/969610 - 137s - loss: 2.8339 - acc: 0.1232 - val_loss: 2.8353 - val_acc: 0.1235
Epoch 50/500
969610/969610 - 138s - loss: 2.8339 - acc: 0.1232 - val_loss: 2.8353 - val_acc: 0.1235
Epoch 51/500
969610/969610 - 138s - loss: 2.8339 - acc: 0.1232 - val_loss: 2.8353 - val_acc: 0.1235
Epoch 52/500
969610/969610 - 138s - loss: 2.8339 - acc: 0.1232 - val_loss: 2.8353 - val_acc: 0.1235
Epoch 53/500
969610/969610 - 139s - loss: 2.8339 - acc: 0.1232 - val_loss: 2.8353 - val_acc: 0.1235
Epoch 54/500
969610/969610 - 138s - loss: 2.8339 - acc: 0.1232 - val_loss: 2.8353 - val_acc: 0.1235
Epoch 55/500
969610/969610 - 138s - loss: 2.8339 - acc: 0.1232 - val_loss: 2.8353 - val_acc: 0.1235
Epoch 56/500
969610/969610 - 138s - loss: 2.8338 - acc: 0.1232 - val_loss: 2.8353 - val_acc: 0.1235
Epoch 57/500
969610/969610 - 138s - loss: 2.8338 - acc: 0.1232 - val_loss: 2.8352 - val_acc: 0.1235
Epoch 58/500
969610/969610 - 137s - loss: 2.8338 - acc: 0.1232 - val_loss: 2.8352 - val_acc: 0.1235
Epoch 59/500
969610/969610 - 138s - loss: 2.8338 - acc: 0.1232 - val_loss: 2.8352 - val_acc: 0.1235
Epoch 60/500
969610/969610 - 138s - loss: 2.8338 - acc: 0.1232 - val_loss: 2.8352 - val_acc: 0.1235
Epoch 61/500
969610/969610 - 138s - loss: 2.8338 - acc: 0.1232 - val_loss: 2.8352 - val_acc: 0.1235
Epoch 62/500
969610/969610 - 138s - loss: 2.8338 - acc: 0.1232 - val_loss: 2.8352 - val_acc: 0.1235
Epoch 63/500
969610/969610 - 138s - loss: 2.8338 - acc: 0.1232 - val_loss: 2.8352 - val_acc: 0.1235
Epoch 64/500
969610/969610 - 138s - loss: 2.8338 - acc: 0.1232 - val_loss: 2.8352 - val_acc: 0.1235
Epoch 65/500
969610/969610 - 139s - loss: 2.8338 - acc: 0.1232 - val_loss: 2.8352 - val_acc: 0.1235
Epoch 66/500
969610/969610 - 139s - loss: 2.8338 - acc: 0.1232 - val_loss: 2.8352 - val_acc: 0.1235
Epoch 67/500
969610/969610 - 139s - loss: 2.8338 - acc: 0.1232 - val_loss: 2.8352 - val_acc: 0.1235
Epoch 68/500
969610/969610 - 138s - loss: 2.8338 - acc: 0.1232 - val_loss: 2.8352 - val_acc: 0.1235
Epoch 69/500
969610/969610 - 139s - loss: 2.8338 - acc: 0.1232 - val_loss: 2.8352 - val_acc: 0.1235
Epoch 70/500
969610/969610 - 138s - loss: 2.8338 - acc: 0.1232 - val_loss: 2.8352 - val_acc: 0.1235
Epoch 71/500
969610/969610 - 138s - loss: 2.8338 - acc: 0.1232 - val_loss: 2.8352 - val_acc: 0.1235
Epoch 72/500
969610/969610 - 138s - loss: 2.8338 - acc: 0.1232 - val_loss: 2.8352 - val_acc: 0.1235
Epoch 73/500
969610/969610 - 138s - loss: 2.8338 - acc: 0.1232 - val_loss: 2.8352 - val_acc: 0.1235
Epoch 74/500
969610/969610 - 138s - loss: 2.8338 - acc: 0.1232 - val_loss: 2.8352 - val_acc: 0.1235
Epoch 75/500
969610/969610 - 139s - loss: 2.8338 - acc: 0.1232 - val_loss: 2.8352 - val_acc: 0.1235
Epoch 76/500
969610/969610 - 138s - loss: 2.8338 - acc: 0.1232 - val_loss: 2.8352 - val_acc: 0.1235
Epoch 77/500
969610/969610 - 138s - loss: 2.8338 - acc: 0.1232 - val_loss: 2.8352 - val_acc: 0.1235
Epoch 78/500
969610/969610 - 139s - loss: 2.8338 - acc: 0.1232 - val_loss: 2.8352 - val_acc: 0.1235
Epoch 79/500
969610/969610 - 138s - loss: 2.8338 - acc: 0.1232 - val_loss: 2.8352 - val_acc: 0.1235
Epoch 80/500
969610/969610 - 138s - loss: 2.8338 - acc: 0.1232 - val_loss: 2.8352 - val_acc: 0.1235
Epoch 81/500
969610/969610 - 138s - loss: 2.8338 - acc: 0.1232 - val_loss: 2.8352 - val_acc: 0.1235
Epoch 82/500
969610/969610 - 138s - loss: 2.8338 - acc: 0.1232 - val_loss: 2.8352 - val_acc: 0.1235
Epoch 83/500
969610/969610 - 138s - loss: 2.8338 - acc: 0.1232 - val_loss: 2.8352 - val_acc: 0.1235
Epoch 84/500
969610/969610 - 139s - loss: 2.8338 - acc: 0.1232 - val_loss: 2.8352 - val_acc: 0.1235
Epoch 85/500
969610/969610 - 138s - loss: 2.8338 - acc: 0.1232 - val_loss: 2.8352 - val_acc: 0.1235
Epoch 86/500
969610/969610 - 138s - loss: 2.8338 - acc: 0.1232 - val_loss: 2.8352 - val_acc: 0.1235
Epoch 87/500
969610/969610 - 138s - loss: 2.8338 - acc: 0.1232 - val_loss: 2.8352 - val_acc: 0.1235
Epoch 88/500
969610/969610 - 138s - loss: 2.8337 - acc: 0.1232 - val_loss: 2.8351 - val_acc: 0.1235
Epoch 89/500
969610/969610 - 138s - loss: 2.8337 - acc: 0.1232 - val_loss: 2.8352 - val_acc: 0.1235
Epoch 90/500
969610/969610 - 138s - loss: 2.8337 - acc: 0.1232 - val_loss: 2.8351 - val_acc: 0.1235
Epoch 91/500
969610/969610 - 138s - loss: 2.8337 - acc: 0.1232 - val_loss: 2.8351 - val_acc: 0.1235
Epoch 92/500
969610/969610 - 139s - loss: 2.8337 - acc: 0.1232 - val_loss: 2.8351 - val_acc: 0.1235
Epoch 93/500
969610/969610 - 138s - loss: 2.8337 - acc: 0.1232 - val_loss: 2.8351 - val_acc: 0.1235
Epoch 94/500
969610/969610 - 138s - loss: 2.8337 - acc: 0.1232 - val_loss: 2.8351 - val_acc: 0.1235
Epoch 95/500
969610/969610 - 138s - loss: 2.8337 - acc: 0.1232 - val_loss: 2.8351 - val_acc: 0.1235
Epoch 96/500
969610/969610 - 138s - loss: 2.8337 - acc: 0.1232 - val_loss: 2.8351 - val_acc: 0.1235
Epoch 97/500
969610/969610 - 139s - loss: 2.8337 - acc: 0.1232 - val_loss: 2.8351 - val_acc: 0.1235
Epoch 98/500
969610/969610 - 138s - loss: 2.8337 - acc: 0.1232 - val_loss: 2.8351 - val_acc: 0.1235
Epoch 99/500
969610/969610 - 138s - loss: 2.8337 - acc: 0.1232 - val_loss: 2.8351 - val_acc: 0.1235
Epoch 100/500
969610/969610 - 139s - loss: 2.8337 - acc: 0.1232 - val_loss: 2.8351 - val_acc: 0.1235
Epoch 101/500
969610/969610 - 139s - loss: 2.8337 - acc: 0.1232 - val_loss: 2.8351 - val_acc: 0.1235
Epoch 102/500
969610/969610 - 139s - loss: 2.8337 - acc: 0.1232 - val_loss: 2.8351 - val_acc: 0.1235
Epoch 103/500
969610/969610 - 138s - loss: 2.8337 - acc: 0.1232 - val_loss: 2.8351 - val_acc: 0.1235
Epoch 104/500
969610/969610 - 138s - loss: 2.8337 - acc: 0.1232 - val_loss: 2.8351 - val_acc: 0.1235
Epoch 105/500
969610/969610 - 138s - loss: 2.8337 - acc: 0.1232 - val_loss: 2.8351 - val_acc: 0.1235
Epoch 106/500
969610/969610 - 138s - loss: 2.8337 - acc: 0.1232 - val_loss: 2.8351 - val_acc: 0.1235
Epoch 107/500
969610/969610 - 138s - loss: 2.8337 - acc: 0.1232 - val_loss: 2.8351 - val_acc: 0.1235
Epoch 108/500
969610/969610 - 138s - loss: 2.8337 - acc: 0.1232 - val_loss: 2.8351 - val_acc: 0.1235
Epoch 109/500
969610/969610 - 138s - loss: 2.8337 - acc: 0.1232 - val_loss: 2.8351 - val_acc: 0.1235
Epoch 110/500
969610/969610 - 138s - loss: 2.8337 - acc: 0.1232 - val_loss: 2.8351 - val_acc: 0.1235
Epoch 111/500
969610/969610 - 139s - loss: 2.8337 - acc: 0.1232 - val_loss: 2.8351 - val_acc: 0.1235
Epoch 112/500
969610/969610 - 138s - loss: 2.8337 - acc: 0.1232 - val_loss: 2.8351 - val_acc: 0.1235
Epoch 113/500
969610/969610 - 138s - loss: 2.8337 - acc: 0.1232 - val_loss: 2.8351 - val_acc: 0.1235
Epoch 114/500
969610/969610 - 138s - loss: 2.8337 - acc: 0.1232 - val_loss: 2.8351 - val_acc: 0.1235
Epoch 115/500
969610/969610 - 138s - loss: 2.8337 - acc: 0.1232 - val_loss: 2.8351 - val_acc: 0.1235
Epoch 116/500
然后我决定手动停止/中断model.fit（），然后运行时自动终止并TF抛出错误：
在处理上述异常期间，发生了另一个异常：
Traceback (most recent call last):
File "/usr/local/lib/python3.6/dist-packages/IPython/core/interactiveshell.py", line 1823, in showtraceback
stb = value._render_traceback_()
AttributeError: 'TypeError' object has no attribute '_render_traceback_'
During handling of the above exception, another exception occurred:
Traceback (most recent call last):
File "/usr/local/lib/python3.6/dist-packages/IPython/core/ultratb.py", line 1132, in get_records
return _fixed_getinnerframes(etb, number_of_lines_of_context, tb_offset)
File "/usr/local/lib/python3.6/dist-packages/IPython/core/ultratb.py", line 313, in wrapped
return f(*args, **kwargs)
File "/usr/local/lib/python3.6/dist-packages/IPython/core/ultratb.py", line 358, in _fixed_getinnerframes
records = fix_frame_records_filenames(inspect.getinnerframes(etb, context))
File "/usr/lib/python3.6/inspect.py", line 1490, in getinnerframes
frameinfo = (tb.tb_frame,) + getframeinfo(tb, context)
File "/usr/lib/python3.6/inspect.py", line 1448, in getframeinfo
filename = getsourcefile(frame) or getfile(frame)
File "/usr/lib/python3.6/inspect.py", line 696, in getsourcefile
if getattr(getmodule(object, filename), '__loader__', None) is not None:
File "/usr/lib/python3.6/inspect.py", line 733, in getmodule
if ismodule(module) and hasattr(module, '__file__'):
File "/usr/local/lib/python3.6/dist-packages/tensorflow/__init__.py", line 50, in __getattr__
module = self._load()
File "/usr/local/lib/python3.6/dist-packages/tensorflow/__init__.py", line 44, in _load
module = _importlib.import_module(self.__name__)
File "/usr/lib/python3.6/importlib/__init__.py", line 126, in import_module
return _bootstrap._gcd_import(name[level:], package, level)
File "<frozen importlib._bootstrap>", line 994, in _gcd_import
File "<frozen importlib._bootstrap>", line 971, in _find_and_load
File "<frozen importlib._bootstrap>", line 953, in _find_and_load_unlocked
ModuleNotFoundError: No module named 'tensorflow_core.compat'
Traceback (most recent call last):
File "<ipython-input-61-6592e2e5fa00>", line 5, in <module>
callbacks=[reduce_lr] )
File "/usr/local/lib/python3.6/dist-packages/tensorflow_core/python/keras/engine/training.py", line 727, in fit
workers=workers,
File "/usr/local/lib/python3.6/dist-packages/tensorflow_core/python/keras/engine/training_arrays.py", line 675, in fit
File "/usr/local/lib/python3.6/dist-packages/tensorflow_core/python/keras/engine/training_arrays.py", line 394, in model_iteration
if not isinstance(batch_outs, list):
File "/usr/local/lib/python3.6/dist-packages/tensorflow_core/python/keras/backend.py", line 3476, in __call__
"""Generates a callable that runs the graph.
File "/usr/local/lib/python3.6/dist-packages/tensorflow_core/python/client/session.py", line 1472, in __call__
run_metadata_ptr)
KeyboardInterrupt
我在做什么错了？
任何帮助将不胜感激。