cluster:usage_tips:mlflow

Differences

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

Link to this comparison view

Next revision		 Previous revision
cluster:usage_tips:mlflow [2022/09/30 08:55] – created lukabcluster:usage_tips:mlflow [2022/11/28 05:48] (current) – [Installing ML Flow] matiasrus
Line 2: Line 2:
  
 How-to-Guide for setting up and connecting to ML Flow, tracking experiments and analysing them. How-to-Guide for setting up and connecting to ML Flow, tracking experiments and analysing them.
 +
 +ML Flow Structure:
 +
 +{{:cluster:usage_tips:mlflow_with_sleepvae.png?600|}}
  
 ==== Installing ML Flow ==== ==== Installing ML Flow ====
Line 14: Line 18:
  
 <code> <code>
-mlflow server --host 0.0.0.0 --port=5001+ARTIFACT_ROOT="file://home/your-username@sleep.ru.is/your-directory" 
 +</code>  
 + 
 +Choose any port number e.g. 5001 (you cannot choose the same port number as other users on the cluster) 
 + 
 +<code> 
 +mlflow server --host 0.0.0.0 --port=5001 --default-artifact-root ARTIFACT_ROOT
 </code>     </code>    
  
Line 21: Line 31:
 4) Create a tunnel from the cluster localhost:5000 to your own localhost port to view and compare the experiments through your own browser: 4) Create a tunnel from the cluster localhost:5000 to your own localhost port to view and compare the experiments through your own browser:
          
-    Run the following code on your computers command prompt to map the localhost ports:+Run the following code on your computers command prompt to map the localhost ports:
          
-    <code> +<code> 
-    ssh -L localhost:5000:130.208.209.2:5001 matiasrus@130.208.209.2 +ssh -L localhost:5000:130.208.209.4:your-portnumber your-username@130.208.209.4 
-    </code>+</code>
          
 5) Open localhost:5000 through your browser 5) Open localhost:5000 through your browser
Line 36: Line 46:
 import mlflow import mlflow
  
-mlflow_tracking_uri = 'http://localhost:5000'+mlflow_tracking_uri = 'http://localhost:your-port'
 mlflow_experiment = 'test' mlflow_experiment = 'test'
  
Line 48: Line 58:
 3) Start the tracking 3) Start the tracking
  
-Track parameters using:+Track parameters by putting these lines in the code of your model:
  
 <code> <code>
-    mlflow.log_param('testparameter', testparam)+mlflow.log_param('testparameter', testparam)
 </code> </code>
  
Line 57: Line 67:
  
 <code> <code>
-    mlflow.log_metric('testparameter', testparam, step=epoch)+mlflow.log_metric('testparameter', testparam, step=epoch)
 </code> </code>
  
 Track artefacts (figures, codes, etc. need to be saved before logging) using: Track artefacts (figures, codes, etc. need to be saved before logging) using:
 <code> <code>
-    mlflow.log_artifact('/tmp/test_fig.png') +mlflow.log_artifact('/tmp/test_fig.png') 
-    </code>+</code>
  • cluster/usage_tips/mlflow.1664528142.txt.gz
  • Last modified: 2022/09/30 08:55
  • by lukab