piperider-python-sdk 0.0.1

PipeRider Python SDK

PipeRider Python SDK

PipeRider SDK is a python library to manipulate your project resources on the PipeRider.

Configuration

Before using the sdk, you need a api key to authenticate with the api server. Visit the user settings page and generate your api key:

• https://app.piperider.io/settings

When the api key available, set it to the environment:

export PIPERIDER_API_KEY=<user_api_key>

Or setup it with help function:

from PipeRider.api.http_client import set_api_key

set_api_key('api-key')

Install

pip install piperider-python-sdk

Usage

First, start a new project my-pipe-rider

import PipeRider

project = PipeRider.project('my-pipe-rider')

Then, create a run to keep ML project settings:

with project.runs.create(name='Awesome Run') as run:
    run.config = {'learning_rate': 0.02,
                  'architecture': 'CNN',
                  'dataset': 'TKNV-users', }

    run.params = {
        'batch_size': 64,
        'epoch': 100,
        'learning_rate': 0.005
    }

Don't forget to link dataset, it is useful when reviewing how a model built:

run.add_dataset('golden-dataset')

The run could be used logging metrics and mark the final result:

offset = random.random() / 5
for ii in range(2, 10):
    acc = 1 - 2 ** -ii - random.random() / ii - offset
    loss = 2 ** -ii + random.random() / ii + offset
    # 2️⃣ Log metrics from your script to PipeRider
    run.log({"acc": acc, "loss": loss})

# the final metrics
run.metrics = {"acc": 0.987, "loss": 0.123}

You also could add comment to the timeline:

project.comment('it is a good idea.')

Finally, mark the run as winner if it was so outstanding:

run.win()