MLBLocks-Demos

Some example MLPipelines and code to test them on sample datasets

Installation

Just run

pip install -r requirements.txt

This will pull and install MLBlocks from the githup repository.

No package installation command is required, as all the scripts are independent.

Special requirements

In order to be able to run the audio pipeline, ffmpeg needs to be installed:

sudo apt-get install ffmpeg

Usage

The simplest way to run an example is by runnnig each script inside the examples folder independently:

python examples/path/to/the/example.py

However, the script run_example.py provides a more convenient way to run them, with the possibility to pass custom arguments and store the results for later review.

It accepts the following parameters:

• example: It can be module path, or the FQN of a module or just the example name. If an example name is given, some default arguments are passed to it.
• --output, -o: path to a folder where the results of the test will be stored as a JSON file.
• extra arguments: Any extra argument written after the example name as key=value is passed to the example run method.

These are some valid examples:

python run_example.py examples/image/simple_cnn_classifier.py
python run_example.py examples.image.simple_cnn_classifier
python run_example.py simple_cnn_classifier test_size=0.3 epochs=12
python run_example.py simple_cnn_classifier -o test_results