AI and data on

Getting started with the NeMo Chainguard Container

Thu, 16 May 2024 08:00:00 +0200

Chainguard’s NeMo container image provides a security-hardened environment for NVIDIA’s NeMo deep learning framework with minimal vulnerabilities compared to traditional AI/ML containers. NeMo enables building conversational AI models through module collections for Automatic Speech Recognition (ASR), Natural Language Processing (NLP), and Text-to-Speech (TTS) tasks. Built for CUDA 12 GPU acceleration, this lightweight container maintains full NeMo functionality while significantly reducing security risks for both training and production inference workloads.

What is Deep Learning?

Deep learning is a subset of machine learning that leverages a flexible computational architecture, the neural network, to address a wide variety of tasks. Neural networks emulate the structure of the brain and consist of interconnected nodes (neurons) that each contain an associated weight and threshold. In concert with an activation function, these values determine whether data is propagated within the network, producing an output layer corresponding to a classification, regression, or other result.

Getting started with the PyTorch Chainguard Container

Thu, 25 Apr 2024 08:00:00 +0200

Chainguard’s PyTorch container image provides a security-hardened foundation for deep learning workloads. Built with PyTorch and CUDA support for GPU acceleration, this minimal image maintains full deep learning capabilities while reducing attack surface. This guide demonstrates fine-tuning models and secure inference deployment.

What is Deep Learning?

Getting started with the Chainguard Spark FIPS container

Thu, 04 Jun 2026 00:00:00 +0000

Apache Spark is a distributed computing engine for batch processing, stream processing, and machine learning at scale. Organizations subject to federal compliance requirements—including FedRAMP, FISMA, and Department of Defense frameworks—must use FIPS 140-3 validated cryptography for all cryptographic operations in Spark.

Chainguard’s Spark FIPS container packages Apache Spark with the Bouncy Castle FIPS cryptographic provider, replacing the standard JVM cryptographic modules with NIST-validated equivalents. In FIPS mode, TLS connections require BCFKS-format keystores rather than the standard PKCS12 or JKS formats, and only FIPS-approved cipher suites are permitted.