In this post, we will discuss the new liquid neural networks and what they might mean for the vision field


Excitement is building in the artificial intelligence community around MIT’s recent release of liquid neural networks. The breakthroughs that Hasani and team have made are incredible.

Let’s dive in.

YouTube version of this post.

What are Liquid Neural Networks?

The Backdrop

Artificial intelligence research and applications involve the construction and training of deep neural networks. Until liquid neural networks, all deep learning systems have shared the same vulnerability — namely, that they learn a fixed mapping from input data to output prediction based on the training data that they are shown, making them brittle to the shifting environment around them. Furthermore, most deep learning models are context…

Earlier this week, OpenAI dropped a bomb on the computer vision world — you can now make image classifications with no training required.


Earlier this week, OpenAI dropped a bomb on the computer vision world: two new groundbreaking models that hint at what’s to come as massive GPT3-esque Transformer models encroach on the vision domain. While DALL-E (a model that can generate images from text prompts) has garnered much of the attention this week, this post focuses on CLIP: a zero-shot classifier which is arguably even more consequential.

Until now, classifying images has involved collecting a custom dataset of hundreds, thousands, or even millions of labeled images that suitably represent your targeted classes and using it to train a supervised classification model (usually…

Practical approaches to processing text with deep learning, explained

Natural language processing (NLP) is a branch of artificial intelligence that encompasses a wide area of software designed to reason about and act on text data. NLP technology is rapidly advancing and it can be difficult to sort out which NLP techniques are making the most impact in industry.

In this post, we will introduce and discuss the natural language processing techniques that have become widely adopted and used in real life applications.

Let’s dive in.

A Brief History of Natural Language Processing

The field of natural language processing traces its roots back all the way to Alan Turing and the Turing test — can a computer…

Object detection technology advances with the release of Scaled-YOLOv4. This blog is written to help you apply Scaled-YOLOv4 to your custom object detection task, to detect any object in the world, given the right training data.

Improve your YOLOv4 training flows with these tips — and please note: we have also published YOLOv4 tactics on our blog.


The YOLO v4 model is currently one of the best architectures to use to train a custom object detector, and the capabilities of the Darknet repository are vast. In this post, we discuss and implement ten advanced tactics in YOLO v4 so you can build the best object detection model from your custom dataset.

Note: this discussion assumes that you have already trained YOLO v4. To get started, check out our guide on training YOLOv4. The advanced tactics here will build from that basic training flow.

YOLO v4 Advanced Tactics RoadMap:

  • Gather More Data
  • Image Preprocessing and Augmentation
  • Image Input…

Note: We have also posted What is a Label Map on our blog. In this post, we will demystify the label map by discussing the role that it plays in the computer vision annotation process. Then we will get hands on with some real life examples using a label map.

Image by author

In computer vision datasets, the question is pervasive — what is a label map?

In this post, we will demystify the label map by discussing the role that it plays in the computer vision annotation process. Then we will get hands on with some real life examples using a label map.

Common Errors Requiring a Label Map

  • Class names missing
  • Class names show as integers
  • Class names do not match my dataset’s class names
  • Class labels not recognized
  • Class labels are numbers

The Role of the Label Map

Computer vision datasets come in all flavors of formats. Roboflow supports the injection, conversion, and exportation of over 30 computer vision formats. …

Note: we have also published Tackling the Small Object Problem on our blog. Detecting small objects is one of the most challenging and important problems in computer vision. In this post, we will discuss some of the strategies we have developed at Roboflow by iterating on hundreds of small object detection models.

Small objects as seen from above by drone in the public aerial maritime dataset

To improve your model’s performance on small objects, we recommend the following techniques:

If you prefer video, I have also recorded a discussion of this post

Why is the Small Object Problem Hard?

The small object problem plagues object detection models worldwide. Not buying it? Check the COCO evaluation results for recent state of the art models YOLOv3, EfficientDet, and YOLOv4:

Note: We have also published PP-YOLO Surpasses YOLOv4 on our blog. Baidu publishes PP-YOLO and pushes the state of the art in object detection research.

PP-YOLO evaluation shows faster inference (x-axis) with better accuracy (y-axis)

PP-YOLO evaluation metrics show improved performance over YOLOv4, the incumbent state of the art object detection model. Yet, the Baidu authors write:

This paper is not intended to introduce a novel object detector. It is more like a recipe, which tell you how to build a better detector step by step.

Let’s unpack that.

YOLO Development History

YOLO was originally authored by Joseph Redmon to detect objects. Object detection is a computer vision technique that localizes and tags objects by drawing a bounding box around them and identifying the class label that a given box belongs too. …

Note: We have also published Train TensorFlow 2 Object Detection model on our blog. Learn how to install, load custom data, train, and infer with your custom TensorFlow 2 Object Detection model to detect any object in the world.

With the recent release of the TensorFlow 2 Object Detection API, it has never been easier to train and deploy custom state of the art object detection models with TensorFlow. To build a custom model you can leverage your own custom dataset to detect your own custom objects: foods, pets, mechanical parts, and more.

In this blog and TensorFlow 2 Object Detection Colab Notebook, we walk through how you can train your own custom object detector in minutes, by changing a single line of code for your dataset import.

Train your custom object detector with the TensorFlow2 Object Detection API

In order to train our custom object detector with the TensorFlow…

Note: We have also published How to Train Detectron2 on our blog. In this post, we will walk through how to train Detectron2 to detect custom objects in this Detectron2 Colab notebook. After reading, you will be able to train your custom Detectron2 detector by changing only one line of code for your custom data import!

Detectron2 provides a flexible framework to train and deploy computer vision algorithms.

Article Outline

  • Overview of Detectron2
  • Overview of our custom dataset
  • Install Detectron2 dependencies
  • Download custom Detectron2 object detection data
  • Visualize Detectron2 training data
  • Write our Detectron2 training configuration
  • Run Detectron2 training
  • Evaluate Detectron2 performance
  • Run Detectron2 inference on test images

Custom Detectron2 Training Resources

Overview of Detectron2

Detectron2 is a popular PyTorch based modular computer vision model library. It is the second iteration of Detectron, originally written in Caffe2. The Detectron2 system allows you to plug in custom state of the art computer vision technologies into your workflow. Quoting the Detectron2 release blog:

Detectron2 includes all the…

Jacob Solawetz

Writes about forward deployed AI in computer vision — machine learning engineer @

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store