This paper was published by Mingxing Tan and Quoc V. Le, members of Google Research, Brain Team. They propose a method to scale up neural nets. Their biggest model achieves **84.4% top-1 / 97.1% top-5 accuracy on ImageNet** and the state-of-the-art on CIFAR-100, while being **8.4x smaller and 6.1x faster on inference** than the best existing ConvNet.

EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks (May 2019)

This paper introduces several tricks that can significantly improve the accuracy of common models. And those tricks are quite easy to implement. Here are my notes.

Bag of Tricks for Image Classification with Convolutional Neural Networks (Dec 2018)

Recently I had to convert a python code base into `Cython` binaries. Because we are using `poetry`, it was not straight forward. Here is the process to do it step by step. 

Using Cython with Poetry

Set up the cache in your Github Actions workflows and poetry.

Setting up the cache on Github Actions and Poetry

In this article, I'm sharing my method to read 1 paper per day.

My Research Paper Reading Routine in 3 steps

In the last article I talked about the course I took at Coursera. This time I'm going to talk about another specialization: [Advanced Machine Learning with TensorFlow on Google Cloud Platform]([https://www.coursera.org/specializations/advanced-machine-learning-tensorflow-gcp).

Coursera Specialization "Advanced Machine Learning with TensorFlow on Google Cloud Platform"

Last January I started the Coursera specialization [Architecting with Google Compute Engine Specialization](https://www.coursera.org/specializations/gcp-architecture). In this article, I'm going to share my experience and information about this course.

Coursera Specialization "Architecting with Google Compute Engine"

Old paper but since there was some breakthrough in self-supervised learning, I'm studying deeper this topic recently.

Multi-task Self-Supervised Visual Learning (Aug 2017)

Great paper! Very pragmatic and it works well!

nnU-Net: Self-adapting Framework for U-Net-Based Medical Image Segmentation (Sep 2018)

Other notes that I took from the papers I read for my project in Medical Image Segmentation. It's not a variation of UNet.

Deep Contextual Networks for Neuronal Structure Segmentation (Feb 2016)

In this article, I'm going to cover the usage of TensorFlow 2 and tf.data on a popular semantic segmentation 2D images dataset: ADE20K. This code is now runnable on colab.

Semantic Segmentation with tf.data in TensorFlow 2 and ADE20K dataset

I'm currently working on a biomedical 2D image segmentation project so I'm reading all related papers. I took some notes and here is the first of the series. UNet, a very popular architecture.

U-Net: Convolutional Networks for Biomedical Image Segmentation (May 2015)

Short note on how to remove the error messages like `Module 'cv2' has no '...' member` on VS Code.

Solving 'Module ... has no member...' in VS Code

I'm continuing to take notes about my mistakes/difficulties using TensorFlow. I had Keras ImageDataGenerator that I wanted to wrap as a tf.data.Dataset. I couldn't adapt the documentation to my use case. Here is a concrete example for image classification.

Keras ImageDataGenerator and tf.Data.Dataset in TensorFlow 2.0

I had some trouble using TensorFlow 2.0 with my GPU without using Docker. Sometimes my cuda version is not compatible with the TensorFlow build, other times it's about cudnn ... Using Anaconda makes your life easier!

Installing TensorFlow with Anaconda for your GPU

A common limitation or pain point in a project involving machine learning is the amount of annotated data needed to train our models. In this paper, the authors are exploring what they call Contrastive Predictive Coding. This method allows their models to learn representations (features) in an unsupervised manner. In the end, the goal is to provide a training process that would require much less annotated data.

Representation Learning with Contrastive Predictive Coding (Jan 2019)

In this article, we are going to set up an NVIDIA Jetson Nano.

All Posts

EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks (May 2019)

Bag of Tricks for Image Classification with Convolutional Neural Networks (Dec 2018)