# TFLite
Source: [https://docs.qualcomm.com/doc/80-70014-15B/topic/tflite.html](https://docs.qualcomm.com/doc/80-70014-15B/topic/tflite.html)
TensorFlow Lite (TFLite) is an open-source deep learning framework for on-device inference.
TFLite helps developers run their models on mobile, embedded, and edge platforms by
optimizing the model for latency, model size, power consumption, etc. Qualcomm supports
executing TFLite models natively on Qualcomm Linux hardware through TFLite Delegates as
listed below.
| Delegate | Acceleration |
| --- | --- |
| AI Engine Direct Delegate (QNN Delegate) | CPU, GPU and HTP |
| XNNPack Delegate | CPU |
| GPU Delegate | GPU |
Reference Guide
[Reference Guide](https://docs.qualcomm.com/bundle/publicresource/topics/80-70014-54/overview.html)
Run Sample App
[Run](https://docs.qualcomm.com/bundle/publicresource/topics/80-70014-50/ai-ml-sample-applications.html)
API Reference
[C/C++](https://www.tensorflow.org/lite/api_docs/cc?_gl=1*xotull*_ga*OTM5NzE4ODUwLjE2ODk2OTQ3MDI.*_ga_W0YLR4190T*MTcxNTM3NzQyMS4yNy4xLjE3MTUzNzc0MjEuMC4wLjA)
Last Published: Jan 21, 2026
[Previous Topic
AI Hub](https://docs.qualcomm.com/bundle/publicresource/80-70014-15B/topics/ai-hub.md) [Next Topic
Qualcomm Neural Processing Engine](https://docs.qualcomm.com/bundle/publicresource/80-70014-15B/topics/snpe.md)