Day 1: As part of #100DaysOfCode

- I created a twitting bot for posting tweets from CLI, less distraction and can still track my learning progress.
- Added @intelsoftware OpenVINO Toolkit docker based image for portability: https://t.co/WTRpfMzE18#100DaysOfCode #Code pic.twitter.com/WHHkqICnsU

— Mpho Mphego (@MphoMphego) April 17, 2020

Day 2 and 3:
- Worked my serverless daily diary capstone final project for @udacity Cloud Developer nano degree.
- Made some improvements on the tweeter bot to handle multiple user accounts.
- Improved my @intelsoftware OpenVINO docker image.'#100DaysOfCode #Code

— Mpho Mphego (@MphoMphego) April 19, 2020

Day 4:
- No Code day...
- Watched a talk @anjuan gave at @devconfza titled: 'Managing the Burnout Burndown'#100DaysOfCode #Code

— Mpho Mphego (@MphoMphego) April 20, 2020

Day 5:
1. Take initiative and never stop learning.
2. Teach others what you know.
3. Assume nothing, question everything.
4. Respect constructive criticism.
5. Give credit where it's due.#100DaysOfCode #Code

— Mpho Mphego (@MphoMphego) April 22, 2020

Day 6:
- Spent sometime debugging my serverless application offline using: https://t.co/LNOeATX72B
- Dockerized my dev env with help from: https://t.co/6Yu4cmu6BR#100DaysOfCode #Code

— Mpho Mphego (@MphoMphego) April 22, 2020

Day 7:
- Updated the serverless application and ensured that the rest API works with the correct endpoints using postman.
- Upgrade from Ubuntu 18.04 to 20.04, which took more time than I anticipated.#100DaysOfCode #Code

— Mpho Mphego (@MphoMphego) April 24, 2020

Day 8:
- Finally submitted my @udacity #CloudDeveloper capstone project, which is a simple serverless diary application.
I guess now I can call myself a full-stack developer LOL!#100DaysOfCode #Code

— Mpho Mphego (@MphoMphego) April 24, 2020

Day 9:
- Finally completed @udacity #CloudDeveloper Nanodegree program, it;s been a long 3.5 months.

Fork and make a PR to improve my code: https://t.co/uLabF0r8ZC#100DaysOfCode #UdacityLevelUp #Code pic.twitter.com/fzzYqZh6T9

— Mpho Mphego (@MphoMphego) April 25, 2020

Day 9:
- Spend sometime going through the basics of the Intel Distribution of OpenVINO Toolkit.
- Familiarizing myself with different Computer Vision model types and choosing the right Pre-Trained Model for your App#100DaysOfCode #Code #UdacityLevelUp

— Mpho Mphego (@MphoMphego) April 26, 2020

Day 10 and 11:
- Watched @lexfridman's Can Machines Think? Alan Turing's paper discussion video.
- Spent today explaining the basics of radio astronomy, vlbi and dft to our jnr engineer.#100DaysOfCode #Code #UdacityLevelUp

— Mpho Mphego (@MphoMphego) April 28, 2020

Day 12:
Didn't code much, but spent a lot of time watching PyCon2020 talks.
- Static Typing in Python by @di_codes was interesting.
- Automate the Boring Stuff with Slackbo by @takanory#100DaysOfCode #Code #UdacityLevelUp

— Mpho Mphego (@MphoMphego) April 30, 2020

Day 13/100:

- I recently upgraded my Dell XPS 15 to #Ubuntu 20.04 LTS, and decided to write a blog post with some troubleshooting steps.https://t.co/huHzDoD0L3#100DaysOfCode #Code #UdacityLevelUp

— Mpho Mphego (@MphoMphego) April 30, 2020

Day 14&15:
- Finished a lesson on, The Model Optimizer as part of the @udacity #Intel Edge AI Nanodegree program.
- Worked through some exercizes and updated docker images.#100DaysOfCode #Code #UdacityLevelUp

— Mpho Mphego (@MphoMphego) May 3, 2020

Day 16:
- Documented ways on converting Tensorflow, ONNX, Caffe models to Intermediate Representation which is used with the Inference Engine on the @IntelSoftware OpenVINO Toolkit.

Collaborations are welcome: https://t.co/4nSTqU1Bkq#100DaysOfCode #Code #UdacityLevelUp

— Mpho Mphego (@MphoMphego) May 5, 2020

Day 17:
- Spent half of my day trying to automate SOCKS5 proxy authentication on Selenium Firefox geckodriver for a webscrapper. This isn't working out, I guess I'll have to resort to using firefox profiles and modify creds before loading#100DaysOfCode #Code #UdacityLevelUp

— Mpho Mphego (@MphoMphego) May 6, 2020

Day 18:
Spent sometime explaining TDD and why we need it to our intern and went through an example on out #Python Automated Qualification Framework.#100DaysOfCode #Code #UdacityLevelUp

— Mpho Mphego (@MphoMphego) May 8, 2020

Day 19:
I spent most of my day working on my latest blog post on, My experience and what I learned from @Udacity Cloud Developer Nanodegree.https://t.co/hA1tFeRHLY#100DaysOfCode #Code #UdacityLevelUp

— Mpho Mphego (@MphoMphego) May 8, 2020

Day 20:
Battled with vehicle detection which reports the color and type of vehicle through OpenCV and intel openVINO.#100DaysOfCode #Code #UdacityLevelUp

— Mpho Mphego (@MphoMphego) May 12, 2020

Day 21:
- Fixed my vehicle detection code using OpenCV and OpenVINO by choosing the right model and optimizing the code.
- Complete lesson on deploying an edge app, streaming video over the net and doing inference on it#100DaysOfCode #Code #UdacityLevelUp

— Mpho Mphego (@MphoMphego) May 14, 2020

Day 22:
- Went from the basics of AI at the Edge, built my skills with pre-trained models, the Model Optimizer, Inference Engine and more with the @Intelsoftware OpenVINO TL, and even learned more about deploying an app at the edge.#100DaysOfCode #Code #UdacityLevelUp

— Mpho Mphego (@MphoMphego) May 14, 2020

Day 23:
- Researched on possible models to use for @udacity @intelsoftware Edge AI person counter project.
- Updated documentation on using docker image to download models from model zoo and converting TF model to Intermidiate Representation#100DaysOfCode #Code #UdacityLevelUp

— Mpho Mphego (@MphoMphego) May 16, 2020

Day 24:
- Finally managed to run OpenCV on docker container with x11, but now am having issues with running inference on my frames.
- Updated some documentation wrt to running GUI apps on docker#100DaysOfCode #Code #UdacityLevelUp

— Mpho Mphego (@MphoMphego) May 17, 2020

Day 25:
- Got side tracked from my learning path, had to update a webscrapper to support SOCKS5 proxy auth and cloudflare spoofing with alternating Useragent, So many tabs to close.#100DaysOfCode #Code #UdacityLevelUp

— Mpho Mphego (@MphoMphego) May 17, 2020

Day 26 & 28:
- Decided to use SSD MobileNet COCO model with @IntelSoftware #OpenVINO to count people moving in/out of ROI for my @udacity project.
- Code still flaky but kinda works, will spend next few days to fix it.#100DaysOfCode #Code #OpenCV #UdacityLevelUp pic.twitter.com/nshNrmTzyc

— Mpho Mphego (@MphoMphego) May 19, 2020

Day 29 & 30:
- Spent the past 2.5 days trying to, detect when a person crosses the blue line and exiting on the red line. Point of all this is to count number of people in the video using #OpenVINO and #OpenCV2.#100DaysOfCode #Code #UdacityLevelUp pic.twitter.com/V9OulxVWR9

— Mpho Mphego (@MphoMphego) May 21, 2020

Day 31 & 33:
- Finally completed my person counter project using #OpenVINO, encountered few glitches but it came through.
- Sadly I failed to run it on docker containers though but mmmyyeeh!#100DaysOfCode #Code pic.twitter.com/Qo1h8lC2Nz

— Mpho Mphego (@MphoMphego) May 24, 2020

Day 34:
- Working on a blog post documenting my person counter project.
- Planning to revive my @youtube channel with a walk-through on my project.#100DaysOfCode #Code #UdacityLevelUp

— Mpho Mphego (@MphoMphego) May 25, 2020

Day 35 & 36:
- Compiled and Published a post on, "AI At The Edge - An Introduction To @IntelSoftware Intel #OpenVINO Toolkit."https://t.co/luTsW6iy66

- A sequel to this post will be on the project I completed recently: Person Counter App.#100DaysOfCode #Code #UdacityLevelUp

— Mpho Mphego (@MphoMphego) May 28, 2020

Day 37&41:

- Compiled and Published a post title: "Face Mask Detection Using Intel #OpenVINO And #OpenCV"https://t.co/kPvhIWbzF4
- Created my First @YouTube tutorial (My English kept running away and I butchered it lol.)https://t.co/N4XTA898Ll#100DaysOfCode #UdacityLevelUp

— Mpho Mphego (@MphoMphego) June 2, 2020

Day 42:
- Reviewed and used the Intel DevCloud for running deep learning models on the CPU and integrated GPU.#100DaysOfCode #Code #UdacityLevelUp

— Mpho Mphego (@MphoMphego) June 7, 2020

Day 43:
- Made progress wrt @udacity EdgeAI, lessons on how to choose the right hardware and inference model, based on a number of factors.#100DaysOfCode #Code #UdacityLevelUp

— Mpho Mphego (@MphoMphego) June 8, 2020

Day 44:
- Re-watched one of my fav talks by @hillelogram
- Currently listening to @testandcode episode: 107 (Property Based Testing in Python with Hypothesis - Alexander Hultner)https://t.co/CulNpLeTC4https://t.co/LlQLRFIxEx#100DaysOfCode #Code #UdacityLevelUp

— Mpho Mphego (@MphoMphego) June 9, 2020

Day 45:
- Researched ways of running inference on an FPGA (Intel Vision Accelerator Design (VAD) with Intel Arria FPGA) using the @intelsoftware DevCloud with pre-existing bitstreams ensuring that they match my network.#100DaysOfCode #Code #UdacityLevelUp

— Mpho Mphego (@MphoMphego) June 11, 2020

Day 46-48:
- I've been working on a "Smart Queue Monitoring System" which is running on various hardware on @intelsoftware DevCloud.
- Currently works 100% on CPU, next it's GPU, VPU then finally an FPGA.

Time now: 2:50 AM Phew!!!!#100DaysOfCode #Code #UdacityLevelUp pic.twitter.com/HuNR0chzyi

— Mpho Mphego (@MphoMphego) June 13, 2020

Day 49-52:
- Finally completed my "Smart Queue Monitoring System" project which was tested on various hardware (CPU, GPU, VPU and FPGA.) on @intelsoftware DevCloud using @IntelSoftware #OpenVINO.

- Link to repo and study in the comment.#100DaysOfCode #Code #UdacityLevelUp

— Mpho Mphego (@MphoMphego) June 17, 2020

Day 53:
- Learned about Generative AI (GANs) and @awscloud DeepComposer.
- Built a custom Generative Adversarial Network (GAN) to generate music using DeepComposer on @awscloud SageMaker.#100DaysOfCode #Code #UdacityLevelUp

— Mpho Mphego (@MphoMphego) June 20, 2020

Day 54-58:

- Spent the week investigating ways of running @micropython on an ESP32-CAM for video streaming purposes.
- The idea is to stream video from esp32 to a Raspberry Pi running @IntelSoftware OpenVINO toolkit.https://t.co/ym7xucjpNl#100DaysOfCode #Code #UdacityLevelUp

— Mpho Mphego (@MphoMphego) June 26, 2020

Day 59-60:

- Got derailed and worked on my #EasyEquities side project that scrapes my portfolio and updates a Google sheet with my current investments.
- Made a shareable sheet for anyone to use (Assuming you in RSA)https://t.co/gPSUJpMf9u#100DaysOfCode #Code #UdacityLevelUp

— Mpho Mphego (@MphoMphego) July 1, 2020

Day 61:
- Integrated #GoogleSheet API with my #EasyEquities scrapper to push data directly to my #GoogleSheet, and added script as cronjob every 2 hrs during business hours.

Now I can easily track my wins and losses!https://t.co/6hT94Cbb4V#100DaysOfCode #Code #UdacityLevelUp pic.twitter.com/zagg7IZ1H6

— Mpho Mphego (@MphoMphego) July 1, 2020

Day 62:

- Learned about Optimization Techniques and tools relevant to edge computing.
- Learned how to measure the performance of models by using FLOPs and MACs as performance metric#100DaysOfCode #Code #UdacityLevelUp

— Mpho Mphego (@MphoMphego) July 3, 2020

Day 63:
- Learned that when optimising a model consider separable (depthwise and pointwise) convolutional layer over the standard convolutional layer, by doing so the number of FLOPs(floating point operations) to run the model is reduced.#100DaysOfCode #Code #UdacityLevelUp

— Mpho Mphego (@MphoMphego) July 5, 2020

Day 64:
- I learned about the different techniques for benchmarking model size and reduction, by means of quantisation, weight sharing and knowledge distillation.
- Played around with the @intelsoftware OpenVINO DL Workbench.#100DaysOfCode #Code #UdacityLevelUp

— Mpho Mphego (@MphoMphego) July 6, 2020

Day 65:
- Today was all about advanced application profiling using the @intelsoftware VTune Amplifier and finding bottlenecks in my application. I do not like installing foreign software in my PC so I will spend few hours dockerizing it.#100DaysOfCode #Code #UdacityLevelUp

— Mpho Mphego (@MphoMphego) July 7, 2020

Day 66-69:
- Been up since 2AM due to the storm and decided to continue working on my project that uses the @IntelSoftware #OpenVINO #FaceDetection, gaze and other models to control a Computer Mouse Pointer using @AlSweigart's PyAutoGUI module.#100DaysOfCode #EdgeAI #Python pic.twitter.com/y1UIEUXcty

— Mpho Mphego (@MphoMphego) July 13, 2020

Day 70-71:
- Finally completed my blog post that has been in my draft for almost 3 week titled:

How I Deployed A People Counter App At The Edge Using @Intelsoftware OpenVINO.

>>> Link in the comments. <<<#100DaysOfCode #EdgeAI #Python pic.twitter.com/8SVftSWYkn

— Mpho Mphego (@MphoMphego) July 18, 2020

Day 71-74:
- Completed my project that uses the @IntelSoftware
#OpenVINO #FaceDetection, #GazeEstimation and other models to control a Computer Mouse Pointer using #PyAutoGUI module.
- Blog post and (flaky) code coming soon!#100DaysOfCode #EdgeAI #Python pic.twitter.com/0PtlrfQSPj

— Mpho Mphego (@MphoMphego) July 18, 2020

Day 75:
- Completed my blog post titled: How I Deployed A Computer Pointer Controller Using Gaze Estimation.

- Graduated from the @udacity @IntelSoftware #EdgeAI for IoT Developers #nanodegree
#100DaysOfCode #EdgeAI #Python

— Mpho Mphego (@MphoMphego) July 20, 2020

Day 76:
- Revisited my #COVID19 #FaceMask detection project and improved it slightly.
- Performance increased, total model load an inference time on both models decreased.
- Only running mask inference if face is detected.https://t.co/rnw1eM8SuK
#100DaysOfCode #EdgeAI #Python

— Mpho Mphego (@MphoMphego) July 22, 2020

Day 77:
- Revisiting Computer Vision fundamentals from understanding the importance of light source and the maths behind CV using Matlab (I know!!)

This is through a CV basics course from @coursera #100DaysOfCode #EdgeAI #Python

— Mpho Mphego (@MphoMphego) July 26, 2020

Day 78:
- Going back to the basics, learned about the 3-levels paradigm of computer vision (cognitive science as David Marr's Tri-Level Hypothesis).
- Computational theory, Representation & Algorithms, and Implementation#100DaysOfCode #EdgeAI #Python

— Mpho Mphego (@MphoMphego) July 31, 2020

Day 79:
- Been meaning to create a REST API for one of my project and have been gravitating towards #fastapi for a while now.

After watching the talk+demo from @tiangolo on @abhi1thakur's channel - I am sold.#100DaysOfCode #EdgeAI #Python

— Mpho Mphego (@MphoMphego) August 1, 2020

Day 80:
- Back on my path of AI+CV
- I've started looking at #skimage lib for manipulating images as opposed to #OpenCV, and from where I stand seems well-structured. Will need to check the performance between #skimage and OpenCV.#100DaysOfCode #EdgeAI #Python #ImageProcessing

— Mpho Mphego (@MphoMphego) August 2, 2020

Day 81-83:
- Learned how to do restore images in order to remove objects, logos, text, or damaged areas in images!
- Also learned how to denoise, segmentation, and find elements in images by their contours using #SkikitImage#100DaysOfCode #EdgeAI #Python #ImageProcessing

— Mpho Mphego (@MphoMphego) August 5, 2020

Day 83:
- Enhanced the contrast of this image so that it may appear more dramatic using the Contrast Limited Adaptive Histogram Equalization (CLAHE)#100DaysOfCode #EdgeAI #Python #ImageProcessingDatacamp pic.twitter.com/026fRarvPq

— Mpho Mphego (@MphoMphego) August 5, 2020

Day 84:

- Learned about image denoising, restoration and most importantly segmentation using the SLIC (Simple Linear Iterative Clustering) filter algorithm which creates Superpixels based on k-mean clustering. #100DaysOfCode #EdgeAI #Python #ImageProcessingDatacamp pic.twitter.com/MJT2Rmg8Oy

— Mpho Mphego (@MphoMphego) August 9, 2020

Day 85:

- Learned about some useful feature extraction mainly edge (Canny) and corner (Harris Corner) detection (very useful for object detection etc).#100DaysOfCode #EdgeAI #Python #ImageProcessingDatacamp

— Mpho Mphego (@MphoMphego) August 11, 2020

Day 86-89:

- Completed the Image Processing with Python course on @DataCamp
- Rewatched talks given by the Python Sensei @dontusethiscode
-- https://t.co/B7MxQxf3bT
-- https://t.co/OrFHB98FIh#100DaysOfCode #EdgeAI #Python #ImageProcessingDatacamp

— Mpho Mphego (@MphoMphego) August 17, 2020

Day 87-90:
Juggling between a number of things is one of the traits I've learned to master (I choose to believe that bhahaha)

- Halfway into week 3, re-learning about NNs and DLs from @coursera's @deeplearningai_ course offered by @AndrewYNg#100DaysOfCode #Python #DeepLearning pic.twitter.com/TO6rXndp3N

— Mpho Mphego (@MphoMphego) August 18, 2020

Day 91:
- Compiled a blog post titled: Face Blurring Using Scikit-image Cascade Classifiers.

- Based on some of the skills learned from @DataCamp's Image Processing with Python course.

<< LINK IN THE COMMENT >>#100DaysOfCode #Python #DeepLearning #ImageProcessingDatacamp

— Mpho Mphego (@MphoMphego) August 20, 2020

Day 92-94:
- Completed all lessons for Neural Networks and Deep Learning @deeplearningai
- Working on a gesture controller project that uses facial landmarks to control game play for my son as part of his exercise routine#100DaysOfCode #Python #DeepLearning

— Mpho Mphego (@MphoMphego) August 27, 2020

Day 96:
- Working on gesture control project which uses facial landmarks and tracks a persons nose then determine the body direction of movement.
- My son loves "Subway Surfer" , so I thought he would love playing it using body movements.#100DaysOfCode #Python #DeepLearning pic.twitter.com/GrrtXdey3B

— Mpho Mphego (@MphoMphego) August 29, 2020

Day 97-125+:
- Finalizing my game controller using body movements.
- Wrote 2 blogs: How to create face detection in under 10 lines and adding makefiles to your projects
- Created a #OpenVINO utility library and a youtube video demoing it.#100DaysOfCode #Python #DeepLearning

— Mpho Mphego (@MphoMphego) September 24, 2020