# LSTM on Micro-Controller

**URL:** <https://discuss.tinyml.seas.harvard.edu/t/lstm-on-micro-controller/681>\
**Category:** General Discussions\
**Created:** [May 11, 2021, 5:33am UTC](https://discuss.tinyml.seas.harvard.edu/t/lstm-on-micro-controller/681 "2021-05-11T05:33:17Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![Predstan](https://avatars.discourse-cdn.com/v4/letter/p/f14d63/32.png) [@Predstan](https://discuss.tinyml.seas.harvard.edu/u/Predstan)\
**Post date:** [May 11, 2021, 5:33am UTC](https://discuss.tinyml.seas.harvard.edu/t/lstm-on-micro-controller/681/1 "2021-05-11T05:33:17Z")

</div>

Here is a project that i just completed. Inspiration from the TinyML class and Pete Warden’s codes. I built a new model from scratch that works well for this application. But i have one question. Can you run LSTM layers on micro-controller. I am hoping to expand this project that i just completed with LSTM or GRU layers. Is there an available literature or paper that can help with the LSTM implentation on edge devices

[![](https://canada1.discourse-cdn.com/flex027/uploads/tinyml1/original/1X/c5d4602e8f3b6cc031b099ffc02ecdb7af35e97f.jpeg "Speech Commands for Home appliances") ](https://www.youtube.com/watch?v=deWhHhF_Atk)

---

<div class="post-metadata">

**Author:** ![vjreddi](https://yyz1.discourse-cdn.com/flex027/user_avatar/discuss.tinyml.seas.harvard.edu/vjreddi/32/35_2.png) [@vjreddi](https://discuss.tinyml.seas.harvard.edu/u/vjreddi)\
**Post date:** [May 11, 2021, 7:55pm UTC](https://discuss.tinyml.seas.harvard.edu/t/lstm-on-micro-controller/681/2 "2021-05-11T19:55:27Z")

</div>

Cool, and to answer your question about LSTMs… no we can’t run LSTMs or GRU layers net in TFLM.