The following content is displayed on the same screen for your experiment so that you can make any necessary reference in experiment. Start your experiment now!

1. Login to the deep learning server via SSH using PuTTY. (Automatically done for you)(Duration: 2 mins)
   This lab will be entirely based on operations in PuTTY
2. Check installation package information,Obtain the location of dlicmd.py(Duration: 5 mins)
   

   ($EGO_TOP/dli/1.2.2/dlpd/bin)
   Execute the following command to make sure the command file exists.

   EGO_TOP=/opt/ibm/cluster/wmla123

   source $EGO_TOP/profile.platform

   ls -l $EGO_TOP/dli/1.2.5/dlpd/bin/dlicmd.py

3. Obtain Master Host Information(Duration: 3 mins)
   

   Execute the following commands line by line
   egosh user logon
   （user accout: please use the assigned os user, password is same with username）
   
   MASTER_HOST=`hostname -f`

   echo $MASTER_HOST

4. Prepare the environment for using dlicmd.py,Source Anaconda environment(Duration: 2 mins)
   

   source /opt/anaconda3/etc/profile.d/conda.sh

   conda activate dlipy3

5. Login(Duration: 2 mins)
   

   python $EGO_TOP/dli/1.2.5/dlpd/bin/dlicmd.py --logon --master-host $MASTER_HOST

   （username: please use the assigned os user, password is same with username）
   

6. List all currently available machine learning and deep learning frameworks(Duration: 2 mins)
   

   python $EGO_TOP/dli/1.2.5/dlpd/bin/dlicmd.py --dl-frameworks --master-host $MASTER_HOST
   
   

7. Submit a TensorFlow Single Host job,exec-start to submit(Duration: 3 mins)
   

   python $EGO_TOP/dli/1.2.5/dlpd/bin/dlicmd.py --exec-start PyTorch --master-host $MASTER_HOST --ig sig01 --cs-datastore-meta type=fs,data_path=msdtool/ --model-main /opt/ibm/spectrumcomputing/examples/wmla/BYOF_models/pytorch_examples/linear_regression.py

   

   Check the output Exe id and Copy it（Hold left button and select it）

8. exec-get to query(Duration: 3 mins)
   

   export JOB_ID="$JOBID"
   replace $JOBID with the Exec id generated in the Step 7
   python  $EGO_TOP/dli/1.2.5/dlpd/bin/dlicmd.py --exec-get $JOB_ID  --master-host $MASTER_HOST
   

9. Obtain training results(Duration: 3 mins)
   

   export JOB_ID="$JOBID"
   replace $JOBID with the Exec id generated in the Step 7

   python $EGO_TOP/dli/1.2.5/dlpd/bin/dlicmd.py --exec-trainlogs $JOB_ID --master-host $MASTER_HOST
   

Start Your Experiment