Shivaraj karkiIntel OpenVino Boost Semantic Segmentation predictionA booster to predict ML model on production line.Jun 30, 2022Jun 30, 2022
Shivaraj karkiSet-up a Multi GPU server from ScratchThis blog guides to setup a fully working multi GPU server from scratch, with detailed explanation and links. A Beast to Boost AI…Mar 18, 2021Mar 18, 2021
Shivaraj karkiDistributed Computing for multi GPU, Deep Neural Networks with Ray and Tensorflow, for Object…Training Deep Neural Networks in distributed environment with multiple GPUs with Object Detection architectures, yolo v3, tensorflow, RayJan 12, 2021Jan 12, 2021
Shivaraj karkiDilated Convolutions for Image Classification with ResNetThis is not Dilated ResNet (DRN), but faster and better than ResNet and DRN, with 94% lesser parameters than ResNetSep 14, 2020Sep 14, 2020
Shivaraj karkiSnapshot Ensemble model for Multi-class Classification in Deep LearningInstead of training N neural networks independently from scratch, we train one neural network and let it converge M times to local minima.Sep 14, 2020Sep 14, 2020
Shivaraj karkiPython, Power of SetReduce Time complexity using python sets, during data preprocessing.Sep 14, 2020Sep 14, 2020
Shivaraj karkiLoss function in CNN doesn’t get stuck in local minima, Saddle point is the culprit.In one of the lecture video of cs231n, Andrej Karpathy said that in optimization problem global minima are not major problem in loss curve…Sep 10, 2020Sep 10, 2020