https://github.com/arrhythmia-detection/esp32tfmicro

Abstracts away all the low-level complexity of TensorFlow Lite Micro and makes it really easy to deploy TFLM models on supported ESP32 boards. By default, this library prints per ops wise execution time via Serial. This makes it really easy to collect metrics for logging.
https://github.com/arrhythmia-detection/esp32tfmicro

easy-esp32-tflm esp32 esp32-s3 platformio-library tensorflow-lite tensorflow-lite-micro tflm-arduino

Last synced: about 2 months ago
JSON representation

Abstracts away all the low-level complexity of TensorFlow Lite Micro and makes it really easy to deploy TFLM models on supported ESP32 boards. By default, this library prints per ops wise execution time via Serial. This makes it really easy to collect metrics for logging.

• Host: GitHub
• URL: https://github.com/arrhythmia-detection/esp32tfmicro
• Owner: arrhythmia-detection
• License: mit
• Created: 2024-10-29T07:30:18.000Z (over 1 year ago)
• Default Branch: main
• Last Pushed: 2026-04-22T15:58:34.000Z (2 months ago)
• Last Synced: 2026-04-26T23:36:35.174Z (about 2 months ago)
• Topics: easy-esp32-tflm, esp32, esp32-s3, platformio-library, tensorflow-lite, tensorflow-lite-micro, tflm-arduino
• Language: C++
• Homepage:
• Size: 9.53 MB
• Stars: 1
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

          
ESP32TFLMWrapper




    


        

            

        

        

            

        

        

            

        

        

            

        

    


 

        

            

        

        

            

        

        


            


            

                

            

            


        





A simple [PlatformIO](https://platformio.org/) Arduino library that

abstracts away all the low level complexity of TensorFlow

Lite Micro and makes it really easy to deploy

TFLM (*T*ensor*F*low *L*ite *M*icro) models on supported ESP32 boards.

On top of that it adds the ability to print

per [ops](https://www.tensorflow.org/api_docs/cc/namespace/tensorflow/ops) wise

execution time (*latency*).

### Ops Execution Time Logging

By default, the library will log per ops wise execution time.

To stop logging the execution time, please add/append the build flag shown below

in your [platformio.ini](platformio.ini) file and build the project.

```shell

    build_flags = -DEXCLUDE_OPS_EXECUTION_TIME_LOGGING

```

### Acknowledgement

This library internally utilizes

[EloquentTinyML](https://github.com/eloquentarduino/EloquentTinyML)

and [tflm_esp32](https://github.com/eloquentarduino/tflm_esp32) developed

and maintained by [eloquentarduino](https://github.com/eloquentarduino).

> [!NOTE]

> For examples on how to use this library, please check out [examples](examples) folder.

 

Copyright © 2024-present 

   Inmoresentum

    and Contributors.