Skip to main content

Featured

I built An Offline ChatBot Companion Raspberry Pi Robot

  This post also has a  YouTube version. Meet Katherine my personalized offline conversational companion chatbot, which can answer questions, give facts on anything, and can remember and comprehend.  I made this chatbot without any large language models. So no-GPT essentially. After years of working on my own conversational software as a hobby, I finally have a prototype I can use to demonstrate. This is Katherine my offline chatbot. I named it after Katherine Johnson - A human computer that worked for NASA. Katherine is powered by my own python code , some public libraries, and local files. I’ve spent a few years learning different strategies for closed looped conversational system. Thanks to Katherine here, I think I’m on the right track. Why did I make Katherine?  Well aside from having my own personalized chatbot, l wanted to be able to have a companion robot that I could talk to no matter the situation. Whether it be a Power outage, or a camping trip, my compani...

Working with the coral dev board mini

 Coral dev board mini

For a comprehensive tutorial go here. Or for a quick read try this article then proceed to the linked pages.


Specs


    Dev Board Mini Data Sheet  


Synopsis 

The dev board mini is a minimal single board computer. The operating system is not GUI-based so if you’re used to visually clicking files and folders and don’t use the command line at all, you’re in for a real treat.


The coral dev board is designed to be mobile in nature, having no keyboard and mouse connection sockets, and requiring software to be downloaded via terminal on a host computer. This means you can’t even setup the board if you don’t have a spare computer available that can run the mdt tools. Also installing regular python packages seemed to give me trouble, so I ended up researching different ways to get packages I’d already installed on my other single board computer. 


To even connect to the board you must either ssh or physically connect the host and coral board mini via cords, and wait for the device to show up using the command “mdt devices”. In order to execute that command you’ll have already installed the MDT tools for the coral.


Setup:


    Materials 

        Coral dev board mini

        python3 installed linux or mac host computer

        2 usb-c to usb cords

        Wifi

        Patience


To setup the google coral dev board mini, you plug the mini into the host computer for storage via usb-c then also plug a power usb-c cord up to it as well. 




IMPORTANT:

    If you can't start your board or it's acting funny , you've probably bricked it. But that's ok because you can jump start / reset the board by connecting two points together on the board as shown below.


Force-boot into fastboot mode





Note: I strongly advise creating a file containing all the steps and packages just in case you brick it!


Running models


Mozilla DeepSpeech

My experience is specific to the project I’m currently working on (an offline end to end chatbot). I typically use Mozilla DeepSpeech for my speech recognition and espeak for the text to speech component. Both packages require dependencies though and that’s where things can get a little crazy.


In order to install the dependencies I had to google each error that popped up in the command output and try installing the software again. A combination of “apt-get install packageNameHere”and “pip3 install packageNameHere “with the right versions of the programs eventually allowed me to reach my goal. 


What I meant by versioning is that if you’re trying to install a program named B that requires program A version 1.12.14 and You already installed program A version 2 then you’d have to delete and reinstall a specific version in order for it to work. If you’re not paying attention to the command output, the errors can be frustrating and cause you to give up. I specifically had issues with numpy and scipy. 


Performance 

    -Pending



Conclusion 

To sum it all up, if you don’t mind not having a GUI and are good with commands and pip, you should be fine buying this single board computer for any demos or projects that you're ready to test out.

Comments

Popular Posts