Faisal Z. Qureshi
faisal.qureshi@ontariotechu.ca
Here, I collect code samples that some of you may find useful as you develop sophisticated deep learning systems. The over-arching theme of this exercise is to demonstrate that deep learning systems all follow a similar template. Any deep learning system can be devided into the following parts:
- Data;
- Model;
- Compute environment;
- Training; and
- Evaluation.
The first two parts, i.e., Data and Model, are very problem specific, whereas the last three parts can be easily reused between projects.