9. Glossary

Apache TVM

An open source machine learning compiler framework for CPUs, GPUs, and other devices. It enables running and optimization of machine learning computations on such hardware.

compute

A mathematical formula that specifies what an operator does.

Edge AI Studio

TI EdgeAI Studio is a collection of tools that enable development, benchmarking and deployment of AI applications.

inference

Inferencing is the act of using a trained deep learning network to output a prediction about incoming data.

Model Composer

Model Composer is a tool in Edge AI Studio. It is a fully integrated no-code solution for training and compiling models for deployment into edge AI applications. It supports Bring-Your-Own-Data (BYOD), enabling the re-training of models from the TI Model Zoo with custom data. Deploy to your local development board.

NPU

Running layers in the model on the hardware NPU accelerator.

PyTorch

PyTorch is a commonly-used open-source deep learning framework.

QAT

Quantization-Aware Training (QAT) is a quantization technique for mitigating model accuracy degradation that arises from quantization. QAT models the effects of quantization during training, allowing it to achieve the highest accuracy among quantization techniques. Please see the documentation of PyTorch native QAT for more details.