GundamBen opened a new issue #20147:
URL: https://github.com/apache/incubator-mxnet/issues/20147
## Description
Getting error that mxnet does not support SaparableConv1d with mxnet 1.5.0
and CUDA 10.1. This error will not occur when running the same code in
tensorflow 1.14.0.
### Error Message
Traceback (most recent call last):
File "crashModel.py", line 24, in <module>
model = Lstm()
File "crashModel.py", line 17, in Lstm
model.add(SeparableConv1D(filters = 64, kernel_size = (1)))
File
"/root/anaconda3/envs/mxnet/lib/python3.6/site-packages/keras/engine/sequential.py",
line 181, in add
output_tensor = layer(self.outputs[0])
File
"/root/anaconda3/envs/mxnet/lib/python3.6/site-packages/keras/engine/base_layer.py",
line 470, in __call__
output = self.call(inputs, **kwargs)
File
"/root/anaconda3/envs/mxnet/lib/python3.6/site-packages/keras/layers/convolutional.py",
line 1393, in call
dilation_rate=self.dilation_rate)
File
"/root/anaconda3/envs/mxnet/lib/python3.6/site-packages/keras/backend/mxnet_backend.py",
line 3773, in separable_conv1d
raise NotImplementedError('MXNet Backend: Separable Conv1D not supported
yet!')
NotImplementedError: MXNet Backend: Separable Conv1D not supported yet!
## To Reproduce
import os
os.environ["CUDA_VISIBLE_DEVICES"] = "2"
os.environ["KERAS_BACKEND"] = 'mxnet'
from keras.models import Sequential
from keras.layers import Dense, Dropout, Embedding, LSTM, Bidirectional,
Reshape, Input, SeparableConv1D, RepeatVector
def Lstm():
model = Sequential()
model.add(Embedding(input_dim = 140, output_dim = 128))
model.add(Bidirectional(LSTM(128)))
model.add(Dropout(0.5))
model.add(Dense(64, activation='relu'))
model.add(Dropout(0.5))
model.add(RepeatVector(3))
model.add(SeparableConv1D(filters = 64, kernel_size = (1)))
model.add(Reshape((192, )))
model.add(Dense(64, activation='relu'))
model.add(Dense(2))
return model
model = Lstm()
model.summary()
### Steps to reproduce
(Paste the commands you ran that produced the error.)
1. create a python file and paste code above in it.
2. set os.environ["KERAS_BACKEND"] = tensorflow and run the file in
tensorflow, the model works well.
3. set os.environ["KERAS_BACKEND"] = mxnet and run the file in mxnet,
programme reports an error.
## What have you tried to solve it?
1. I used to run the file in the version of 1.5.0. When the error occured, I
updated mxnet to 1.8.0, but it didn't worked.
## Environment
<details>
<summary>Environment Information</summary>
----------System Info----------
Platform : Linux-4.15.0-123-generic-x86_64-with-debian-stretch-sid
system : Linux
node : 3bb941e39736
release : 4.15.0-123-generic
version : #126-Ubuntu SMP Wed Oct 21 09:40:11 UTC 2020
----------Python Info----------
Version : 3.6.12
Compiler : GCC 7.3.0
Build : ('default', 'Sep 8 2020 23:10:56')
Arch : ('64bit', '')
mxnet version 1.8.0
cuda version 10.1
tensorflow version 1.14.0
</details>
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
