Deep Learning with Python using Tensorflow and Keras on AMD GPU with ROCm gives errors when I run the program below

Question

I have an AMD GPU pc which I have recently started to use with Linux Mint OS I've seen a way of installing ROCm on this pc following this tutorial, but when I tried to write a Python program using Keras it threw these errors. Here is my code:

import numpy as np
from numpy import genfromtxt
from sklearn.model_selection import train_test_split
from sklearn.preprocessing import MinMaxScaler
from keras.models import Sequential
from keras.layers import Dense

data=genfromtxt('../DATA/bank_note_data.txt',delimiter=',')
labels=data[:,4]
features=data[:,0:4]
X=features
y=labels 

X_train, X_test, y_train, y_test = train_test_split(X,y,test_size=0.33,random_state=42)

scalerObject=MinMaxScaler()
scalerObject.fit(X_train)
scaled_X_train=scalerObject.transform(X_train)
scaled_X_test=scalerObject.transform(X_test)

model=Sequential()
model.add(Dense(4,input_dim=4,activation='relu'))
model.add(Dense(8,activation='relu'))
model.add(Dense(1,activation='sigmoid'))

model.compile(loss='binary_crossentropy',optimizer='adam',metrics=['accuracy'])

model.fit(scaled_X_train,y_train,epochs=50,verbose=2)
print(model.fit(scaled_X_train,y_train,epochs=50,verbose=2))

These were the errors:

Traceback (most recent call last):
  File "/home/cemosambora/.local/lib/python3.6/site-packages/tensorflow/python/pywrap_tensorflow.py", line 58, in <module>
    from tensorflow.python.pywrap_tensorflow_internal import *
  File "/home/cemosambora/.local/lib/python3.6/site-packages/tensorflow/python/pywrap_tensorflow_internal.py", line 28, in <module>
    _pywrap_tensorflow_internal = swig_import_helper()
  File "/home/cemosambora/.local/lib/python3.6/site-packages/tensorflow/python/pywrap_tensorflow_internal.py", line 24, in swig_import_helper
    _mod = imp.load_module('_pywrap_tensorflow_internal', fp, pathname, description)
  File "/usr/lib/python3.6/imp.py", line 243, in load_module
    return load_dynamic(name, filename, file)
  File "/usr/lib/python3.6/imp.py", line 343, in load_dynamic
    return _load(spec)
ImportError: librccl.so: cannot open shared object file: No such file or directory

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "/home/cemosambora/eclipse-workspace/Deep_Learning_OpenCV/kerasBasics.py", line 5, in <module>
    from keras.models import Sequential
  File "/home/cemosambora/.local/lib/python3.6/site-packages/keras/__init__.py", line 3, in <module>
    from . import utils
  File "/home/cemosambora/.local/lib/python3.6/site-packages/keras/utils/__init__.py", line 6, in <module>
    from . import conv_utils
  File "/home/cemosambora/.local/lib/python3.6/site-packages/keras/utils/conv_utils.py", line 9, in <module>
    from .. import backend as K
  File "/home/cemosambora/.local/lib/python3.6/site-packages/keras/backend/__init__.py", line 89, in <module>
    from .tensorflow_backend import *
  File "/home/cemosambora/.local/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py", line 5, in <module>
    import tensorflow as tf
  File "/home/cemosambora/.local/lib/python3.6/site-packages/tensorflow/__init__.py", line 28, in <module>
    from tensorflow.python import pywrap_tensorflow  # pylint: disable=unused-import
  File "/home/cemosambora/.local/lib/python3.6/site-packages/tensorflow/python/__init__.py", line 49, in <module>
    from tensorflow.python import pywrap_tensorflow
  File "/home/cemosambora/.local/lib/python3.6/site-packages/tensorflow/python/pywrap_tensorflow.py", line 74, in <module>
    raise ImportError(msg)
ImportError: Traceback (most recent call last):
  File "/home/cemosambora/.local/lib/python3.6/site-packages/tensorflow/python/pywrap_tensorflow.py", line 58, in <module>
    from tensorflow.python.pywrap_tensorflow_internal import *
  File "/home/cemosambora/.local/lib/python3.6/site-packages/tensorflow/python/pywrap_tensorflow_internal.py", line 28, in <module>
    _pywrap_tensorflow_internal = swig_import_helper()
  File "/home/cemosambora/.local/lib/python3.6/site-packages/tensorflow/python/pywrap_tensorflow_internal.py", line 24, in swig_import_helper
    _mod = imp.load_module('_pywrap_tensorflow_internal', fp, pathname, description)
  File "/usr/lib/python3.6/imp.py", line 243, in load_module
    return load_dynamic(name, filename, file)
  File "/usr/lib/python3.6/imp.py", line 343, in load_dynamic
    return _load(spec)
ImportError: librccl.so: cannot open shared object file: No such file or directory


Failed to load the native TensorFlow runtime.

This page has some common reasons and solutions.

Answer 1

This error was caused by missing rccl library. After installing rocm, you need to install some ROCm Libraries :

sudo apt-get update && sudo apt-get install -y --allow-unauthenticated  rocm-dkms rocm-dev rocm-libs rccl rocm-device-libs hsa-ext-rocr-dev hsakmt-roct-dev hsa-rocr-dev rocm-opencl rocm-opencl-dev   rocm-utils  rocm-profiler cxlactivitylogger miopen-hip miopengemm

Answer 2

Bit late to respond, but here's my two-cents for what it may be worth. It is highly recommended that we use Ubuntu 18.04 per the pre-requisites and official support offered by ROCm Considering the holistic aspects of the way many eco-system oriented support that you might bump onto in this regard vis-a-vis Python Packages, GPU Support etc., this should be the best bet for you to consider; should you be serious about your endeavour in this regard.