Groq supports standard machine learning frameworks like PyTorch, TensorFlow, and ONNX for inference, but does not currently support ML training with the LPU Inference Engine. The GroqWare suite, including Groq Compiler, provides a push-button experience for custom model development. Users can optimize workloads by hand coding to the Groq architecture and have fine-grained control over GroqChip processors for developing custom applications with maximum performance.

Cover Image for Groq