Pytorch - Zero to GANs: Course Project
As the last assignment to finish the course, we had to choose an online dataset and apply the concepts learned during the course.
1. Finding and choosing a dataset
To start with I was considering using either one related to sports (most likely the Champion's League matches one) or one about e-sports (one about the MOBA DOTA2). However, I ended up choosing this one: . The main reason was that I had previously used a similar one for a mobile app demo I did ~1-2 years ago, so it was interesting to me.
2. Understanding the dataset
Understand and describe the modeling objective clearly: a. What type of data is it? (images, text, audio etc.) b. What type of problem is it? (regression, classification, generative modeling, etc.)
3. Exploratory data analysis
explore the data by plotting graphs and answer any questions you may have
Modeling - try 4-5 approaches
a. Define a model (network architecture) b. Pick some hyperparameters c. Train the model d. Make predictions on samples e. Evaluate on test dataset f. Save the model weights g. Record the metrics
Summarize your learning & identify opportunities for future work.
Link to the Jupyter notebook of my course project: remusa/course-project | Jovian