Differences

This shows you the differences between two versions of the page.

Link to this comparison view

Both sides previous revision Previous revision
Next revision		 Previous revision
doku:deep_learning [2020/04/23 18:32] – [DL and DataLAB community] irdoku:deep_learning [2020/04/24 10:37] (current) – [Deep Learning (DL)] ir
Line 1: Line 1:
 ====== Deep Learning (DL) ====== ====== Deep Learning (DL) ======
  
-under construction 
  
 ===== DL and DataLAB community ===== ===== DL and DataLAB community =====
Line 7: Line 6:
 The [[https://colab.tuwien.ac.at/display/DLDC/|DLDC platform]] is designed to foster the exchange of scientists from different fields in DL. There you will find community meeting announcements and the slides of past meetings. In addition, it is a space to host The [[https://colab.tuwien.ac.at/display/DLDC/|DLDC platform]] is designed to foster the exchange of scientists from different fields in DL. There you will find community meeting announcements and the slides of past meetings. In addition, it is a space to host
 moderated special interest groups in the field of DL. moderated special interest groups in the field of DL.
 +
 +===== DL modules on VSC-3  =====
 +
 +==== Tensorflow ====
 +Typically on some GPU node,
 +<code>
 +[someone@n372-003 ~]$  module purge 
 +[someone@n372-003 ~]$  module load gcc/5.3 mpioff/1.0 bazel/0.12.0 cuda/9.0.176 intel-mkl/2018 hdf5/1.8.18-SERIAL python/3.6 numpy/1.14.3 h5py/2.7.0 scipy/0.18.0 tensorflow/1.8.0.ttest 
 +</code>
 +
 +==== Keras ====
 +Typically on some GPU node,
 +<code>
 +[someone@n372-003 ~]$  module purge 
 +[someone@n372-003 ~]$  module load gcc/5.3 mpioff/1.0 bazel/0.12.0 cuda/9.0.176 intel-mkl/2018 hdf5/1.8.18-SERIAL python/3.6 numpy/1.14.3 h5py/2.7.0 scipy/0.18.0 tensorflow/1.8.0.ttest keras/2.1.6.ttest
 +</code>
 +
 +==== Pytorch ====
 +Typically on some GPU node,
 +<code>
 +[someone@n372-003 ~]$  module purge 
 +[someone@n372-003 ~]$  module load gcc/5.3 cuda/9.1.85 intel-mkl/2018 cmake/3.9.6 python/3.6 numpy/1.14.3 pytorch/1.1.0b
 +</code>
 +
 +==== Theano ====
 +Typically on some GPU node,
 +<code>
 +[someone@n372-003 ~]$  module purge 
 +[someone@n372-003 ~]$  module load gcc/5.3 cuda/9.1.85 intel-mkl/11 cmake/3.8.2 python/2.7 numpy/1.12.0 libgpuarray/0.7.6 scipy/0.17.1 Theano/0.9.0
 +[someone@n372-003 ~]$  export MKL_THREADING_LAYER=GNU
 +[someone@n372-003 ~]$  THEANO_FLAGS=device=cuda,floatX=float32 python whatevermyscript.py
 +</code>
 +
 +Many users nowadays prefer conda-based installs of DL software. For example on some GPU node,
 +<code>
 +[someone@n372-003 ~]$  module purge 
 +[someone@n372-003 ~]$  module load anaconda3/5.3.0
 +[someone@n372-003 ~]$  module show anaconda3/5.3.0 
 +[someone@n372-003 ~]$  source /opt/sw/x86_64/glibc-2.17/ivybridge-ep/anaconda3/5.3.0/etc/profile.d/conda.sh
 +[someone@n372-003 ~]$  conda activate base
 +(base) [someone@n372-003 ~]$  conda list   ( just to see what packages are currently around )
 +(base) [someone@n372-003 ~]$  conda search tensorflow
 +(base) [someone@n372-003 ~]$  conda create -n mytf "tensorflow 1.14.0 gpu_py37h4491b45_0"
 +                                 --> y   ( say yes to the suggested list of packages to install )
 +(base) [someone@n372-003 ~]$  conda deactivate    ( to get out of the 'base' environment again )
 +[someone@n372-003 ~]$  conda info --envs
 +                          --> mytf      /home/lv123456/someone/.conda/envs/mytf
 +[someone@n372-003 ~]$  conda activate mytf
 +(mytf) [someone@n372-003 ~]$  python
 +Python 3.7.blablabla
 +>>> import tensorflow as tf    ( lots of FutureWarning... but probably not harmful !? )
 +>>> hello = tf.constant('Hello, TensorFlow!')
 +>>> sess = tf.Session()
 +>>> print(sess.run(hello))
 +>>> <CTRL>D
 +(mytf) [someone@n372-003 ~]$  conda deactivate
 +</code>
 +
 +
 +
  
  • doku/deep_learning.1587666779.txt.gz
  • Last modified: 2020/04/23 18:32
  • by ir