Gautam KumarOptimizing LLM TrainingTL;DR This blog describes when FSDP is not sufficient and when to add Tensor Parallel (TP).Sep 17Sep 17
Gautam KumarMonetary vs Fiscal policyThese are my notes on learning about economics.Jul 14, 2020Jul 14, 2020
Gautam KumarAnatomy of Kubeflow PipelineTL;DR This blog describes how to replace the mysql instance in kubeflow pipeline to AWS managed RDS (relational databases) instance in…May 15, 20201May 15, 20201
Gautam KumarSageMaker Kubernetes OperatorSagemaker operator is an open source project developed at AWS, which helps kubernetes user to submit machine learning and deep learning…Dec 3, 2019Dec 3, 2019
Gautam KumarinDeepLearning-101Serverless Inference Service on AWS Fargate.AWS launched Fargate in 2017, which allows you to run containers without having to manage servers or clusters. In march 2019 AWS also…Jun 3, 20191Jun 3, 20191
Gautam KumarUnderstanding Training output in TensorFlowRecently I was trying to build an application which will perform image recognition and I trained my model using TensorFlow. However I had…Nov 6, 2018Nov 6, 2018
Gautam KumarinDeepLearning-101Kubernetes with AWS EFSRecently I was trying to perform distributed training of ResNet50 model using TensorFlow. I set up EKS (Elastic Kubernetes Service)…Oct 8, 2018Oct 8, 2018