InfoQ Homepage TensorFlow Content on InfoQ
-
New Technique Speeds up Deep-Learning Inference on TensorFlow by 2x
Researchers at North Carolina State University recently presented a paper at the International Conference on Supercomputing (ICS) on their new technique, "deep reuse" (DR), that can speed up inference time for deep-learning neural networks running on TensorFlow by up to 2x, with almost no loss of accuracy.
-
Google Releases Post-Training Integer Quantization for TensorFlow Lite
Google announced new tooling for their TensorFlow Lite deep-learning framework that reduces the size of models and latency of inference. The tool converts a trained model's weights from floating-point representation to 8-bit signed integers. This reduces the memory requirements of the model and allows it to run on hardware without floating-point accelerators and without sacrificing model quality.
-
Google Releases TensorFlow.Text Library for Natural Language Processing
Google released a TensorFlow.Text, a new text-processing library for their TensorFlow deep-learning platform. The library allows several common text pre-processing activities, such as tokenization, to be handled by the TensorFlow graph computation system, improving consistency and portability of deep-learning models for natural-language processing.
-
Google Announces TensorFlow Graphics Library for Unsupervised Deep Learning of Computer Vision Model
At a presentation during Google I/O 2019, Google announced TensorFlow Graphics, a library for building deep neural networks for unsupervised learning tasks in computer vision. The library contains 3D rendering functions written in TensorFlow, as well as tools for learning with non-rectangular mesh-based input data.
-
Google's Cloud TPU V2 and V3 Pods Are Now Publicly Available in Beta
Recently, Google announced that its second- and third-generation Cloud Tensor Processing Units (TPU) Pods — its scalable cloud-based supercomputers with up to 1,000 of its custom TPU — are now publicly available in beta. With these Pods, Machine Learning (ML) researchers, engineers, and data scientists can speed up the time needed to train and deploy machine learning models.
-
Google Scales Weak Supervision to Overcome Labeled Dataset Problem
Google recognizes that the need for labeled data in machine learning (ML) is a significant bottleneck and recently adapted the open-source Snorkel framework to overcome the problem at scale. Google enhanced Snorkel by integrating it with Tensorflow, using the file system for sharing data instead of a database, and creating separate executables for labeling functions.
-
Teaching the Computer to Play the Chrome Dinosaur Game with TensorFlow.js Machine Learning Library
A simple, yet entertaining and useful for educational purposes application of machine learning, was recently made available on Fritz's HeartBeat Medium publication. Google's machine learning TensorFlow.js library is leveraged in the browser to teach the computer to play the Chrome Dinosaur Game.
-
How the Sequence of Characters in a Name Can Predict Race and Ethnicity
Gaurav Sood, Principal Data Scientist at Microsoft, recently spoke at the AnacondaCon 2019 Conference on how to use the sequence of characters in a person's name to predict that person's race and ethnicity, using machine learning techniques.
-
Google Expands ML Kit, Adds Smart Reply and Language Identification
In a recent Android blog post, Google announced the release of two new Natural Language Processing (NLP) features for ML Kit, including Language Identification and Smart Reply. In both cases, Google is providing domain-independent APIs that help developers analyze and generate text, speak and other types of Natural Language text.
-
Google Open-Sources GPipe Library for Faster Training of Large Deep-Learning Models
Google AI is open-sourcing GPipe, a TensorFlow library for accelerating the training of large deep-learning models.
-
Introducing TensorFlow Privacy, a New Machine Learning Library for Protecting Sensitive Data
In a recent blog post, TensorFlow announced TensorFlow Privacy, an open source library that allows researchers and developers to build machine learning models that have strong privacy. Using this library ensures user data are not remembered through the training process based upon strong mathematical guarantees.
-
Deep Learning for Speech Synthesis of Audio from Brain Activity
Research teams use deep learning neural networks to synthesize speech from electrical signals recorded in human brains, to help people with speech challenges.
-
Uber Open-Sources Ludwig Code-Free Deep-Learning Toolkit
Uber Engineering is open-sourcing Ludwig, a deep-learning toolkit that allows users to experiment with a variety of neural network structures without writing code.
-
TensorSpace.js Delivers Neural Network 3D Visualization Framework
TensorSpace.js provides an open source browser-based neural network data visualization framework to complement the growing machine learning landscape by supporting pre-trained models created with TensorFlow.js, Keras, or TensorFlow.
-
Google Open-Sources BERT: A Natural Language Processing Training Technique
In a recent blog post, Google announced they have open-sourced BERT, their state-of-the-art training technique for Natural Language Processing (NLP) . Google has decided to do this, in part, due to a lack of public data sets that are available to developers. In addition, optimizations have been made to Cloud TPUs to reduce the amount of time required for training NLP.