Production Machine Learning Systems Flashcards
1
Q
What is static and dynamic prediction?
A
- Static prediction - caculate prediction in batches, store it to databse and take the results of a prediction from a database on request
- Dynamic prediction - calculate prediction on the fly
2
Q
What is Peakedness and Cardianality and how it is realated to choosing between static and dynamic model serving?
A
Peakedness - how much is data concentrated aroud the mean
Cardinality - number of values in the set
3
Q
What are the options to make your training faster?
A
- Use a more powerful device
- Optimize your input pipeline
- Distributed training
4
Q
What is a distributed training?
A
Splitting your training workload across multiple worker nodes (potentially with multiple GPU each) connected over a network