Expand the capabilities of your Flyte workflows with a range of powerful integrations.
Crunch data however you want. Automatically transition data from one dataframe type to the other using Structured Dataset.
Use HuggingFace dataset as a native Flyte type.
Visualize and explore big tabular datasets.
Use Polars dataframe as a native Flyte type.
Scale your Pandas workflows.
Validate data at every step of your Flyte workflow.
Validate your data with expectations.
Validate dataframe-like objects.
Databases & Data Warehouses
Manage and connect to databases and warehouses seamlessly.
Run intricate analytical queries with DuckDB
Query a BigQuery table.
Query a Snowflake service.
Apply git-like versioning to your SQL databases.
Execute SQL queries as Flyte tasks.
Query a Hive service.
Process and analyze your data with data-crunchers.
Schedule, monitor and orchestrate Databricks jobs.
Transform data in your warehouses with DBT.
Run Spark jobs on ephemeral clusters.
Query an AWS Athena service.
Store, share and manage features for ML models.
Simplify the model training process.
Train ML models on Sagemaker from within Flyte.
Distributed Model Training
Perform distributed model training to speed up the model development process.
Run Dask jobs natively on a Kubernetes cluster.
Run distributed training with an MPI operator.
Connect to Ray cluster to perform distributed model training and hyperparameter tuning.
Run distributed deep learning workflows.
Run distributed TensorFlow training jobs.
Run distributed PyTorch training jobs.
Streamline the model deployment process.
Generate ONNX models from TensorFlow models.
Generate ONNX models from PyTorch models.
ONNX Scikit Learn
Generate ONNX models from Scikit Learn models.
Monitor data and models from within Flyte.
Track your machine learning metrics with MLFlow
Log any kind of data and generate summaries of datasets.
Exercise greater control over Kubernetes resources.
Configure pods for arbitrary workloads.
Run batch tasks on AWS.
Execute Jupyter notebooks.
Build your own integration
Create your own integration and submit it to the Flyte repository.