Load data for machine learning and deep learning

This section covers information about loading data specifically for ML and DL applications. For general information about loading data, see Ingest data into a Databricks lakehouse.

Store files for data loading and model checkpointing

Machine learning applications may need to use shared storage for data loading and model checkpointing. This is particularly important for distributed deep learning.

Databricks provides the Databricks File System (DBFS) for accessing data on a cluster using both Spark and local file APIs.

Load tabular data

You can load tabular machine learning data from tables or files (for example, see Read CSV files). You can convert Apache Spark DataFrames into pandas DataFrames using the PySpark method toPandas(), and then optionally convert to NumPy format using the PySpark method to_numpy().

Prepare data to fine tune large language models

You can prepare your data for fine-tuning open source large language models with Hugging Face Transformers and Hugging Face Datasets.

Prepare data for fine tuning Hugging Face models

Prepare data for distributed training

This section covers three methods for preparing data for distributed training: Mosaic Streaming, Petastorm, and TFRecords.