FiriuzaTransformer for All Data TypesIn 2017 Google published one of the most cited paper called «Attention is all you need». And a new network architecture for NLP task was…5 min read·May 22, 2022----
FiriuzaDeep Neural Networks for Tabular DataSpoiler: TabNet and TabTransformer will be here…7 min read·Jan 25, 2021----
FiriuzaROI pooling vs. ROI alignIn computer vision there are many interesting problems and one of them is object detection and segmentation on image.4 min read·Jan 8, 2020----
FiriuzainAnalytics VidhyaTensorFlow 2.0: tf.data APIIf you remember the time when only Queues were available in TensorFlow 1.x as the data structure for ETL pipeline (Extract / Transform /…3 min read·Dec 18, 2019----
FiriuzaTensorflow: how to train two networks independentlyIn previous post I wrote about training Several GANs where I faced with difficulties toward to Tensorflow: I couldn’t implement training…2 min read·May 8, 2019----
FiriuzaWeakly Supervised Object LocalizationIn practice dealing with some business cases using deep learning we work with datasets that have only image labels and no any information…7 min read·Oct 3, 2018--1--1
FiriuzaOptimizers for Training Neural NetworksTraining network is iterative process where we want to minimize Loss function:7 min read·Jul 28, 2018----
FiriuzaVantage Point Tree: does careful selection of vantage point make sense?Vantage Point Tree is very old algorithm for making fast search within tree. Off course nowadays we have so many tools for working with big…5 min read·Mar 25, 2018----
FiriuzaOverride TensorFlow Backward-PropagationTensorflow is a great tool that works with deep learning. There are a lot of operations that you easily can implement and make good model…4 min read·Feb 5, 2018--4--4