Skip to content
Permalink
Branch: master
Find file Copy path
Find file Copy path
Fetching contributors…
Cannot retrieve contributors at this time
executable file 49 lines (34 sloc) 3.1 KB

100 Days Of Code - Log

Day 0: August 16, 2018 (Image classification using Tensorflow)

This is my first attempt at working with tensorflow and opencv library.

Today's Progress: Worked onloading data using cv2, training a CNN using tensorflow and saving the model on disk.

Thoughts:

  • This is my first attempt to perform image classification on my own.
  • I have done few courses online but they were all on jupyter notebook and pre-built methods. Hence I thought of working on the pipeline to classify images using tensorflow on my own.
  • I really struggled with tensorflow at first. In fact installing tensorflow and opencv took me a little while. However, it felt great once they started working.
  • I looked for a blog or online resources for help.

I referred the following:

Link to work: https://github.com/ankitameht/image_classification

Day 1: August 17, 2018 (Image classification using Tensorflow cont..)

Todays task was to load the saved model and test it on images unseen by the model.

Today's Progress: Worked saving the model on disk, loading it back again and then using it to test new images.

Thoughts:

  • My struggle was at the point where we have to load the variables (parameters) of tensorflow graph and then test them on images. It took me a while to understand the whole flow of tensorflow.
  • I looked for a blog or online resources for help here too. However, very few had clear details.
  • Some online help and some self intuition did the trick.
  • I was super thrilled to finish my work - training and testing with whole pipeline of tesorflow based image classification

I referred the following:

Link to work: https://github.com/ankitameht/image_classification

Day 2: August 18, 2018 (Udacity Course of AI For Trading)

Now that I have finished one project. I was given an opportunity to learn at Udacity. I have joined the course AI for Trading. My current work profile is Machine Learnig Engineer at FactSet Systesm India Pvt Ltd. My regular work involves looking at financial data and working with them to provide business solution. AI For Trading course is my attempt to learn more about financial data and working on them using AI.

Today's Progress: I have finished a project (small but was a good learning experience) called Trading with Momentum.

Thoughts:

  • This projects involves working with time-series data. I was working on time-series data for the first time.
  • The dataframe after loading the timeseries data has datetime as its index.
  • I worked on various pandas libraries which involved working with the time-series data.
  • A very good learning experience when I was building a trading long short portfolio of stocks and finally calculating its returns.
  • Here I learnt to perform t-test as well.

Link to work: https://github.com/ankitameht/AI-For-Trading-Course-by-Udacity

You can’t perform that action at this time.