workflow_record.md

The record of workflow for Common Test I. Multi-Class Classification

Summary of contents:

• Dependencies

1. Data

   1. Data Dimensions
   2. Data Visualization
   3. Class Distribution
   4. Train/Test/Eval Split

2. Model

   1. Choice of Model
   2. Define & Compile the Model

3. Training & Testing

   1. Choose hyperparameters
   2. Fit to Train & Test Datasets
   3. Evaluate on Eval
   4. Fine-tune & Iterate

---

Create new pipenv

Data

Data Dimensions

Model

Define & Compile the Model

Specify loss function (categorical cross-entropy), optimizer, and metrics