Deep learning workflow

With IBM Spectrum Conductor Deep Learning Impact, you can use either the Caffe or TensorFlow framework to run deep learning workloads.

Before starting with IBM Spectrum Conductor Deep Learning Impact, decide on the framework that you want to use. Based on the framework you use, you create a corresponding dataset and training model. Start a training run using the training model, optionally tuning the hyperparameters and monitoring results between training runs. The final trained model can be validated and used to create an inference model. Alternatively, you can export the trained model weight file.