z-lstm

What is it?

z-lstm is a basic implementation to run predictions with LSTM in Orquestra – a platform for performing computations on quantum computers developed by Zapata Computing.

This project is part the series of Quantum and Classical Machine Learning tutorials in Orquestra, developed by the Quantum Machine Learning team in Zapata.

Usage

In order to use z-lstm in your workflow, you need to add it as a resource:

resources:
- name: z-lstm
  type: git
  parameters:
    url: "git@github.com:zapatacomputing/z-lstm.git"
    branch: "master"

and then import in a specific step:

- - name: my-task
    template: template-1
    arguments:
      parameters:
      - param_1: 1
      - resources: [z-lstm]

Once that is done, you can:

use any template from templates/ directory
use tasks which import resource in the python code.

Submitting workflow jobs

To submit a workflow to Orquestra, first login:

qe login -e <email> -s <server>

Then submit with:

qe submit workflow example/lstm-tutorial.yaml

You can check the workflow processing with:

qe get workflow <workflow_id>

You can check logs with:

qe get logs <workflow_id> -s <step_id>

Finally, you can get workflow results with:

qe get workflowresult  <workflow_id>

Plotting results

To plot the results, run:

python examples/plot_lstm.py <workflow result JSON>

Development and contribution

Create an virtual environment:

virtualenv venv
source venv/bin/activate

Install dependencies:

make install

Running tests

Tests are located in src/python/lstm/*_test.py and can be run with:

make test