Which is the smallest SSD model to use in a MCU like ESP-EYE and using tensorflow lite micro? #50656
Labels
comp:micro
Related to TensorFlow Lite Microcontrollers
stat:awaiting tensorflower
Status - Awaiting response from tensorflower
type:others
issues not falling in bug, perfromance, support, build and install or feature
Hi,
I am trying to work with the smallest SSD model of tensorflow, but my problem is that I am working with an ESP-EYE and this MCU has only 4 MB of memory.
All SSD models that I saw in the pretrained models Garden of tensorflow, are too big (more than 4 MB) or they dont have a ".tflite" file.
In this last case, I saw that there is a way to convert the file ".pb" to ".tflite" HERE and a video where explain which parameters you must use VIDEO, but the size of the files increase a lot.
For example, if the size of a ".pb" file is 5 MB, the conversion makes a file of 25 MB and things like this.
Here is a example of the code in python that I use for the conversion:
Soo my questions are:
I found some documentation, but I´m not sure if it works:
Thanks for read and have a good day.
See you XD
The text was updated successfully, but these errors were encountered: