search

KordingLab / Neural_Decoding

A python package that includes many methods for decoding neural activity
BSD 3-Clause "New" or "Revised" License
423 stars 118 forks source link

Graph execution error: Detected at node 'Adam/StatefulPartitionedCall_2' #18

Open rollwan opened 1 year ago

rollwan commented 1 year ago

When I fiting DNN model in the Examples_decoders_hc.ipynb, it comes an error below:

FailedPreconditionError Traceback (most recent call last) Input In [93], in <cell line: 5>() 2 model_dnn=DenseNNDecoder(units=100,dropout=0.25,num_epochs=10) 4 #Fit model ----> 5 model_dnn.fit(X_flat_train,y_train) 7 #Get predictions 8 y_valid_predicted_dnn=model_dnn.predict(X_flat_valid)

File ~/.local/lib/python3.8/site-packages/Neural_Decoding/decoders.py:396, in DenseNNRegression.fit(self, X_flat_train, y_train) 394 model.fit(X_flat_train,y_train,nb_epoch=self.num_epochs,verbose=self.verbose) #Fit the model 395 else: --> 396 model.fit(X_flat_train,y_train,epochs=self.num_epochs,verbose=self.verbose) #Fit the model 397 self.model=model

File ~/.local/lib/python3.8/site-packages/keras/src/utils/traceback_utils.py:70, in filter_traceback..error_handler(*args, **kwargs) 67 filtered_tb = _process_traceback_frames(e.traceback) 68 # To get the full stack trace, call: 69 # tf.debugging.disable_traceback_filtering() ---> 70 raise e.with_traceback(filtered_tb) from None 71 finally: 72 del filtered_tb

File ~/.local/lib/python3.8/site-packages/tensorflow/python/eager/execute.py:53, in quick_execute(op_name, num_outputs, inputs, attrs, ctx, name) 51 try: 52 ctx.ensure_initialized() ---> 53 tensors = pywrap_tfe.TFE_Py_Execute(ctx._handle, device_name, op_name, 54 inputs, attrs, num_outputs) 55 except core._NotOkStatusException as e: 56 if name is not None:

FailedPreconditionError: Graph execution error:

Detected at node 'Adam/StatefulPartitionedCall_2' defined at (most recent call last): File "/usr/lib/python3.8/runpy.py", line 194, in _run_module_as_main return _run_code(code, main_globals, None,

jglaser2 commented 1 year ago

I'm not sure, as this seems to be a keras/tensorflow issue.

One possible suggestion (coming from here: https://stackoverflow.com/questions/71153492/invalid-argument-error-graph-execution-error) is that it may relate to different versions of the Adam optimizer within different versions of tensorflow.

In decoders.py line 392, you could change this to:

opt = Adam()
model.compile(loss='mse',optimizer=opt,metrics=['accuracy'])  #Set loss function and optimizer

And after line 52 add 

from tensorflow.keras.optimizers.legacy import Adam