ML Ops Examples

[amitkaps]

Examples: ML

• [x] Create Level 1 Example with Penguin Dataset
• [x] Create Level 2 Example with Penguin Dataset
• [ ] Create Level 3 Example with Penguin Dataset

Examples: Peeling the Onion

Basic Model from sklearn / TSFresh

• [ ] Create Level 2 Example with Onion Dataset
• [ ] Create Level 3 Example with Onion Dataset
• [ ] Create Level 4 Example with Onion Dataset

Examples: Idli / Dosa

• [ ] Focus on Size - Pruning / Quantisation
• [ ] Focus on Latency - Size Reduction (above), Batching (Multiple Request)

Starting Up

• [ ] Make a CookieCutter for ML Ops Project

Add Datasets

• [x] Palmerpenguins dataset
• [ ] ML Dataset with 10k rows
• [ ] DL Dataset

[anandology]

For Penguins-level2, do the followin improvements:

• [x] do test/train split, print train and test accuracy
• [x] do grid search on max_depth
• [ ] add simple input validations - expect columns to be present and their data types
• [ ] write documentation using mkdocs-material
• [ ] add sample data with handful of rows to the repo, add tests to validate the whole flow, make it part of github workflows
• [ ] when you run train.py, print meaninful context to the console and also generate a markdown report