zackfreedman / somatic Goto Github PK

hey @ZackFreedman , by any chance can you provide the model header file and gesture header file? that would be really helpful

do you think it will run on "NodeMCU ESP8266 ESP-12F WiFi Lolin V3" insted of the tensy 4.0?

When I try to upload the firmware to my teensy, I get:
'compilation terminated.
exit status 1
QStringList: No such file or directory'
I can't seem to find out what qstringlist is or where I can download it from.

Whenver i try to import the code from github and try to run the init.py file
Then it shows the error "No module named 'somatictrainer'" i.e. in the line 4 of init.py file
I am just trying to make this.
Any help regarding this?

Running this on an esp32 should be doable, right? it would reduce the partslist since it could do the bluetooth keyboard and mouse itself, without needing the bluetooth mate.

[EDIT]Placing these here for future reference
https://github.com/T-vK/ESP32-BLE-Mouse
https://github.com/T-vK/ESP32-BLE-Keyboard
[/EDIT]

It would be really helpful if STL formats of these parts could be made available

Hello. I'm only slightly familiar with the project (it's very cool by the way), but what do you think about moving the TensorFlow Lite model from the Teensy to an android device? The Teensy 4 chugs aways at 600 MHz, while many mid-grade Android devices are multicored and atleast double the clock speed (my already 3 year old LG clocks at 1.4 GHz). I realize that a phone's bluetooth can't just decide to act as HID device (without root?), so maybe use a process like this?:

1. use bluetooth to take signals from the Somatic,
2. send to TF lite on android for processing,
3. send the result back the the Somatic,
4. have the Somatic send the result as a HID device to whatever it is paired to

I know you said in the video that the Teensy works fine even when underclocked, but if the tensorflow processing can be moved off device, maybe something smaller than the teensy could be used.

@ZackFreedman The instructable is telling me to print 'Cosmetic plate.stl' which isn't in the repository.

How possible would it be to simply mirror the training data and train a right-handed version?

Nice projects!

@ZackFreedman, I am working in this project. It is so cool, but the code is quite big and difficult to follow.

Can you give a bit or high level info about how the GUI expected the IMU data?

I would like get it working with a different IMU using quaternions...