Analysis of a Trained Deep Learning Model | New Data on the Edge | Episode 5 | Intel Software


Hi. I’m Meghana Rao. And this is the AI from the
Data Center to the Edge video series. In this episode, you will learn
how to analyze a train deep learning module. We also assess how
well the model is doing against preset thresholds. The goal is to iteratively
test the model and tune hyperparameters until we
get a well-trained model. Alternatively, this
step also tells us if we need to switch
to a different network to achieve expected results. Let’s take a closer
look at what’s involved in model analysis. The first step is assessment
of the model based on metrics such as accuracy,
training and inference speeds, and the size of the model. If the model is
not performing well against preset thresholds
for the above metrics, we tune hyperparameters such
as gradient descent, momentum, solver type, and batch size. This, as we learn in the
course, is an iterative process. We repeat training
with the same network until desired
results are obtained, or we choose a different
network and repeat. The course teaches techniques
like confusion matrix, classification report, and
ROC plot for model analysis. Thanks for watching this episode
of AI from the Data Center to the Edge. Make sure to check out the links
to register for the course. You can complete the lecture
and the notebooks listed in the resources for this course
and join me in the next episode to learn more about deploying
a model to the edge.

Leave a Reply

Your email address will not be published. Required fields are marked *