Again, a recent discussion with my colleague about deploying YOLOv4 on Azure Edge device with FPGA. Our team is trying to deploy a YOLOv4 model on Azure Edge device, and they hope to take advantage of the acceleration provided by FPGA. However, the official Azure Edge document states they only support 5 types of DNN models. They would like to know whether YOLO can be supported and why.

Recently I’ve discussed Bayesian Optimization with my friends and it’s application in Machine learning. Just document them here for future references.

In MLOps, hyper-parameter tuning comes to play when we’ve figured out what ML method we need to use to answer a business question, what data we have, and we’ve started training the model. The performance of the model is highly depended on the hyper-parameters of the model. Therefore, to get a “best” model, we have to spend some time tuning the hyper-parameters. There are three major methods of tuning the hyper-parameters: 1. grid-search, 2. random-search, 3. Bayesian optimization. I will skip some methods which haven’t been not widely used.

On-going work

There are always moments when I’m frustrated and lost my confidence. I think it will be a good idea to document these Kudos and shoutouts that have been given by my teammates, managers, and clients.

Some interview questions

What’s the difference between boosting and bagging

Bagging

• attempts to reduce the chance overfitting complex models.
• Bagging, also called bootstrap, is to create subsets from sample with replacement.
• It trains a large number of “strong” learners in parallel
• aim to reduce overfitting from CART
• uses complex base models and tries to “smooth out” their predictions
• However the results will be dominated by strong features

I recently got my AWS driving license with 958/1000.

My personal feeling is, even with enough hands-on experience, one still need to learn how AWS questions are written or how to break them apart. AWS designed their questions to reflect real-world use cases, sometimes without meticulous care of the wording. Therefore, to handle the controversial questions, it’s easier to take it as a common use case, then dig out the common solutions you’ve implemented before going through the options.

Pyspark Databricks Exercise: RDD

• the purpose of this practice is to get a deeper understanding of the properties of RDD.
• we will not talk about what is rdd and what that means. There are plenty of materials online with excellent explainations.

Pyspark Databricks Exercise

The purpose of this series is to provide some exercise of using pyspark on a cloud platform. Pros:

Load library and create a handler for spark with SparkSession

Let’s put a summary of the common Classification evaluation metrics. What they mean and how to use them.

Accuracy

Meaning:

Correct identifications / all examples

Let’s put a summary of the common linear regression evaluation metrics. What they mean and how to use them.

Mean Absolute Error (MAE)

Meaning:

you should expect you predictions are off MAE from the true value.