mlrunx-integrations 0.1.2

MLRunX integrations for popular ML frameworks

MLRunX Integrations

Framework integrations for MLRunX.

Supported Frameworks

• PyTorch Lightning - MLRunXLogger
• HuggingFace Transformers - MLRunXCallback
• Optuna - MLRunXOptunaCallback
• Hydra - MLRunXHydraCallback

Installation

# Install with all integrations
pip install "mlrunx-integrations[all]"

# Or install specific integrations
pip install "mlrunx-integrations[lightning]"
pip install "mlrunx-integrations[transformers]"
pip install "mlrunx-integrations[optuna]"
pip install "mlrunx-integrations[hydra]"

Usage

PyTorch Lightning

from mlrunx_integrations import MLRunXLogger
from pytorch_lightning import Trainer

trainer = Trainer(logger=MLRunXLogger(project="my-project"))
trainer.fit(model)

HuggingFace Transformers

from mlrunx_integrations import MLRunXCallback
from transformers import Trainer

trainer = Trainer(...)
trainer.add_callback(MLRunXCallback(project="my-project"))
trainer.train()

Optuna

from mlrunx_integrations import MLRunXOptunaCallback
import optuna

study = optuna.create_study()
study.optimize(
    objective,
    callbacks=[MLRunXOptunaCallback(project="my-project")]
)