Python

Exploration Log of Contrastive Language-Image Pre-training

Posted on Jan 25, 2021 | 5 minutes | 1031 words

Introduction Last month OpenAI released CLIP - a neural network that learns to map text and images into the same embedding space using contrastive objective and multi-class N-pair loss. TL;DR The key ingredients of the architecture are: [Read More]

🐍 Supporting the Python Ecosystem

Posted on Jan 9, 2021 | 1 minutes | 105 words

TLDR; If you use Python, please consider supporting the Python Ecosystem by donating to the Python Software Foundation. Click Here to visit PSF Donation Page Donation for the PSF From their website, this is what the funds are used for: [Read More]

python opensource

📙 CIFAR-10 Classifiers: Part 7 - Speed up PyTorch hyperparameter search using Ray Tune

Posted on Oct 24, 2020 | 426 minutes | 90530 words

CIFAR10 Classifier: PyTorch + Ray Tune Edition Objective Explore Ray Tune and in the process attempt to to find a better set of hyperparameters that can beat the best test acurracy: [Read More]

CIFAR10 Classifiers ConvNets GPU PyTorch Ray Tune Python

📙 CIFAR-10 Classifiers: Part 6 - Speed up TensorFlow hyperparameter search using Optuna

Posted on Oct 18, 2020 | 4 minutes | 814 words

CIFAR10 Classfier: TensorFlow + Optuna Edition Our objective is similar to the Keras-Tuner and Ray Tune notebooks: Explore Optuna optimization library for hyperparam tuning Find out if we can beat the test accuracy of a hand tuned model 69. [Read More]

CIFAR10 Classifiers ConvNets GPU TensorFlow Optuna Python

📙 CIFAR-10 Classifiers: Part 5 - Speed up TensorFlow hyperparameter search using Ray Tune

Posted on Oct 3, 2020 | 6 minutes | 1108 words

CIFAR10 Classification: TensorFlow + Ray Tune Edition In this notebook, we turn to hyperparameter tuning using Ray Tune. Our objective is to find better hyperparams and network architecture. Although we will not spend too much time on network architecture, you probably will get enough information to try on your own. [Read More]

CIFAR10 Classifiers ConvNets GPU TensorFlow Ray Tune Python

🍻 💊 Number of People Searching Hangover Cure

Posted on Sep 19, 2020 | 3 minutes | 635 words

This is a remake of chartr.co chart on hangover cure google trend. When I started working on this mini dataviz project, tidy tuesdays of sort, I was hoping to complete it in a few hours. [Read More]

chart-remake dataviz Python Julia R

📙 CIFAR-10 Classifiers: Part 4 - Build a Simple Image Classifier using PyTorch Lightning

Posted on Sep 5, 2020 | 8 minutes | 1572 words

This notebook is part of CIFAR-10 Classifiers post, which covers building classifiers using TensorFlow, PyTorch, PyTorch Lightning etc. Part 1: Tensorflow | Part 3: PyTorch CIFAR 10 Classifier: Pytorch Lightning Edition We use Pytorch Lightning in this notebook, it abstracts away all the boiler plate code that we need to add to every pytorch training workflow. [Read More]

CIFAR10 Classifiers ConvNets GPU PyTorch Lightning Python

🚶🏾‍♂️Random Walk in Python and Julia

Posted on Aug 28, 2020 | 3 minutes | 595 words

Random walk, also known as drunkard’s walk, is starting at some position and randomly taking the next step. In its simplest form, you can generate it in one-dimension using the following pseudocode: [Read More]

code not-a-benchmark Python Julia Numba

📙 CIFAR-10 Classifiers: Part 2 - Use Keras Tuner to speed up the hyperparameter search

Posted on Aug 17, 2020 | 8 minutes | 1507 words

Part 1: Tensorflow This notebook is part of CIFAR-10 Classifiers post. In that post, I cover building classifiers using Deep Learning frameworks such as TensorFlow, PyTorch, PyTorch Lightning and making use of several hyperparameter optimization libraries. [Read More]

CIFAR10 Classifiers ConvNets GPU TensorFlow 2.0 Keras Tuner Python

📙 CIFAR-10 Classifiers: Part 1 - Build a Simple Image Classifier using TensorFlow

Posted on Aug 13, 2020 | 10 minutes | 1971 words

This notebook is part of CIFAR-10 Classifiers post, which covers building classifiers using TensorFlow, PyTorch, PyTorch Lightning etc. CIFAR-10 Classifier: TensorFlow 2.x Edition Objective Build a simple CIFAR10 classification model that runs on GPU using following: [Read More]

CIFAR10 Classifiers ConvNets GPU TensorFlow 2.0 Python