Skip to content

Latest commit

 

History

History
82 lines (66 loc) · 2.03 KB

README.md

File metadata and controls

82 lines (66 loc) · 2.03 KB
Raw

ml-iris

Machine learning for Iris dataset

  • Uses TensorFlow to create a Logistic Regression classifier
  • Full ML pipeline with tensorflow backend

Imports raw data from csv, randomizes data, preprocesses data, splits data, trains model, tests model, saves model

Uses TensorBoard to visualize the results

Dependencies

  • TensorFlow
    • Backend for training and model visualization
  • NumPy
    • Data preprocessing and manipulation
  • Pandas
    • Data preprocessing and manipulation
  • Matplotlib
    • Data visualization backend
  • Seaborn
    • Data visualization helper

Usage

python iris.py [-h] [--visual] [--learning_rate LEARNING_RATE]
               [--filename FILENAME] [--stddev STDDEV]

Train/test iris dataset using logistic regression

optional arguments:
 -h, --help            show this help message and exit
 --load                load model rather than train
 --visual              plot data and features prior to load/test
 --learning_rate LEARNING_RATE
                       learning rate for GradientDescentOptimizer
 --filename FILENAME   file to store/load model to/from
 --stddev STDDEV       standard deviation for random_normal init values

Train a new model

  • Save to default filename, use default hyperparameters
python iris.py

Test a model

  • Load from default filename
python iris.py --load

Custom training

  • Use custom hyperparameters, save to custom file, show features before training
python iris.py --visual --learning_rate 0.1 --filename my_model --stddev 0.5

Custom testing

  • Load from custom file
python iris.py --load --filename my_model

Changelog

  • 0.2.1

    • Rework tensorboard and model saving directory

  • 0.2.0

    • Add TensorBoard support
    • Show validation set accuracy during training
    • Implement tf.name_scope for organization of variables

  • 0.1.2

    • Added command line arguments for learning rate and filename
    • Allow save/restore from any file

  • 0.1.1

    • Changed command line argument implementation to use argparse

  • 0.1.0

    • Initial version