So far, I have successfully completed the following steps: I pulled the latest phdrieger/mltk-container-golden-image-cpu:3.5.0 docker image from docker hub to my laptop using Docker for Windows 10 (v20.10.2) (I have tried "phdrieger/mltk-container-golden-image-gpu:3.5.0" a number of times, but for some reason, it does not work for me...) I ran a docker container (single-mode) from DLTK 3.4 in Splunk Enterprise 8.1.2. The __dev__ container in the DLTK is running (Container Model Info: Container Image = "TensorFlow CPU (deprecated)" / GPU runtime = "none" / Cluster target = "docker"). I can access the JupyterLab, TensorBoard, and API URLs without any issue I can connect to the container via the command line as well. As the next step, I attempted to send some sample data from Splunk to the container using the search command: | inputlookup diabetes.csv | fit MLTKContainer response from * algo=binary_nn_classifier epochs=10 mode=stage into MyModel (The sample search command provided in the "binary_nn_classifier.ipynb" file in Stage 1.) I noticed that the line starting with "| fit MLTKContainer" does not work. Any command lines above "| fit MTLKContainer..." work fine. I tried a couple of other Notebooks for TensorFlow 2.0 listed Jupyter Lab in the container without any luck. All have the same issue. I always make sure that the container is running from both Splunk and Docker sides. Please let me know what I might have missed in the setup/configuration. Thank you.
... View more