mlsolidpy 0.0.2

python client for the mlsolid MLOps platform

Mlsolid-py

mlsolid-py is a python client for mlsolid MLOps platform. This python SDK allows you to track experiments and save artifacts to the mlsolid server. With Model Registries, you can easily version your machine learning models, and quickly access them in your production code.

Installation

# Install grpc client code
pip install zeddo123-mlsolid-grpc-python --extra-index-url https://buf.build/gen/python

# Install mlsolid-py module
pip install mlsolidpy

Usage

mlsolid provides an easy way to track your experiment runs:

from mlsolidpy.mlsolid import Mlsolid

client = Mlsolid('localhost:5000')

print('Experiments', client.experiments)

print('Run ', client.run("urbane-wagon"))

with client.start_run('my_experiment') as run:
    run.log({'checkpoint': "path/to/checkpoint"})
    run.log({'batch-size': 23})

    run.log({'mae': 0.2333, 'loss': 100.0})
    run.log({'mae': 0.2000, 'loss': 90})
    run.log({'mae': 0.1134, 'loss': 10})
    run.log({'metrics': [0.2000, 0.333, 0.2223]})

Model Registry example:

client = Mlsolid('localhost:5000')

created = client.create_model_registry('test_registry_1')

if not created:
    print('could not create model registry')
    exit()

run_id = None

with client.start_run('my_experiment') as run:
    run_id = run.run_id
    run.add_plaintext_artifact('./tests/data/plain_text_file.txt')
    run.add_model('./tests/data/mobile_sam.pt')

added = client.add_model('test_registry_1', run_id, 'mobile_sam.pt', ['latest'])
if not added:
    print('could not add model to registry')
    exit()

client.artifact(run_id, 'plain_text_file.txt')

client.tagged_model('test_registry_1', 'latest')