# 1 Introduction

Source: [https://docs.qualcomm.com/doc/80-PT790-993B/topic/introduction_tools.html](https://docs.qualcomm.com/doc/80-PT790-993B/topic/introduction_tools.html)

Qualcomm Cloud AI 100 Platform and Apps SDKs contain applications and tools to help the user in every step of the deep learning inference workflow. The workflow involves taking a trained model through model conversion, network evaluation (performance, accuracy, etc.), and deployment in an end-to-end machine learning (ML) application.

The applications available in the SDKs are categorized as follows:
      
- SDK version
- Card health, resources, and logs
    - Qaic-util
    - Qaic-log
- Network preparation
    - Model preparator
- Network compilation
    - Qaic-exec
        - Network specialization
        - Single device partitioning
        - Custom I/O
        - Multi-Qaic
        - Mixed precision
        - Intermediate layer outputs
        - Custom operator
    - Qaic-qpc
- Qaic-qpc
- Network accuracy analysis
    - Accuracy evaluator
    - Accuracy analyzer
    - Histogram analysis
    - Test vector generation
- Network performance tuning
    - Model configurator
- Network profiler
    - Compiler trace (Opstats)
    - Opstats analyzer
- Network execution
    - Qaic-runner
    - Over subscription
    - ONNXRT
    - Triton server
- Network deployment
    - Qaic device partition
    - Docker
    - Kubernetes
- End-to-end workflow
    - Inference toolkit
- Object detection postprocessing
    - QAic Smart NMS (non-max suppression)
    - Qdetect

**Parent Topic:** https://docs.qualcomm.com/doc/80-PT790-993B/topic/dl_inference_tools_part.html

Last Published: Jul 26, 2023

[Next Topic
SDK version](https://docs.qualcomm.com/bundle/publicresource/80-PT790-993B/topics/SDK-version.md)