This example demonstrates how to run standard Caffe lenet_solver.prototxt example using Batch AI. This recipe is running on a single compute node.
- For demonstration purposes, MNIST dataset and Caffe configuration file will be deployed at Azure File Share;
- Standard output of the job and the model will be stored on Azure File Share;
- MNIST dataset has been preprocessed according to http://caffe.berkeleyvision.org/gathered/examples/mnist.html available here.
- The original Caffe solver and net prototxt files have been modified to take environment variables: AZ_BATCHAI_INPUT_SAMPLE and AZ_BATCHAI_OUTPUT_MODEL, and available here lenet_solver.prototxt and lenet_train_test.prototxt.
- Since prototxt files supports neither command line overloading nor environment variable, we use job preparation task preparation_script.sh to expand the environment variable specified in the files, providing more flexibility of the job setup.
You can find Jupyter Notebook for this sample in Caffe-GPU.ipynb.
Under Construction...
Under construction...
If you have any problems or questions, you can reach the Batch AI team at [email protected] or you can create an issue on GitHub.
We also welcome your contributions of additional sample notebooks, scripts, or other examples of working with Batch AI.