logo_better.png

Speed up

your Deep Learning Experimentation Workflow by

x30 times.

quick_ml_logo_edited.png

                 quick_ml

              "ML for everyone"

              Quicker, Smarter, Easier ML for Professionals & Beginners

quick_ml is a pypi package aimed at  simplifying and rapidly expediting Computer Vision Workflows for professionals as well as beginners (Image Classification tasks. With Upcoming versions, further support will be provided). With maximum optimizations and minimal lines of code, level up your experimentation process through the help of this tool.
For latest updates, Join us @ Telegram
Click Here
Quick ML Introductory Slides
Click Here

Why

quick_ml?

  • Usual experimentation with various deep learning architectures takes days of time and it is quite exhausting and frustrating. With quick_ml, professionals as well as beginners can obtain results within minutes to hours at max.

  • Fast experimentation. That's it. Experimenting what works and what doesn't is tedious and tiresome.

  • No need to train model architectures one by one. Using Models Training Report feature of quick_ml obtain training results of 24 model architectures in a single go at the fastest speed (1 hour - 2 hours).

  • Using & Setting up TPUs is easy with quick_ml

  • With quick_ml, Datasets for TPUs (tfrecords) are easy & fast to create

  • More than 24 models can be trained in one session

  • 700+ lines of codes condensed to 10-15 lines of codes.

Benchmarks

With the power of TPUs and its most efficient utilization through quick_ml, you can speed up your Deep Learning Workflows by atleast 20 times. 

For instance, an image dataset with 4000 images (dimensions -> 192x192), writing its code usually takes 2 hours to 3 hours. In addition to that, the training time required. With quick_ml, you can save your coding as well as training time, i.e., experimenting with a single model in about 5 -7 lines of code & obtaining results under 5 minutes. 

With maximum optimizations, usually training on single session allows you to experiment maximum 4 model architectures. With quick_ml, you can experiment more than 24 models in a single go all within 1 hour for a dataset with properties similar to the mentioned. 

Specifications

Support for Kaggle Notebooks as well as Google Colab Notebooks w/ TPU enabled ON. For best performance, import the pretrained weights dataset in the Kaggle Notebook. (https://www.kaggle.com/antoreepjana/tf-keras-pretrained-model-weights)


Tensorflow version==2.4.0
Python 3.6+

Note -> Don't import tensorflow in the beginning. With the upcoming updates in the tensorflow, it might take some time to reflect the corresponding changes to the package. The package is built and tested on the most stable version of tensorflow mentioned in the Specifications.

Few Words About the package

The idea behind designing the package was to reduce the unncessary training time for Deep Learning Models. The experimentation time if reduced can help the people concerned with the package to focus on the finer details which are often neglected. In addition to this, there are several utility functions provided at a single place aimed at expediting the ML workflow. These utility functions have been designed with ease of use as the foremost priority and attempt has been made to optimize the TPU computation as well as bring in most of the functionalities. Attempt has been made to reduce about 500-700 lines of code or even more (depending on what you are about to use) to less than 10 lines of code. Hope you like it!

Worked Examples

Best way to learn quick_ml is through the help of code examples.

And then refer to the documentation for the parts of code which you didn't understand.

Go to the Worked Examples Section.