Siby Jose PlathottamTesla AI day: Image space to Vector space — the real breakthroughThe media (both Tesla fans and haters), analysts, and content creators are focused on the Tesla bot and DOJO that were revealed on Tesla…Aug 25, 2021Aug 25, 2021
Siby Jose PlathottamTesla D1 chip: Comparison with A100Here is a quick comparison of Tesla's newly unveiled D1 chip and NVIDIA A100 GPU. Both can be used for training neural networks.Aug 20, 2021Aug 20, 2021
Siby Jose PlathottaminDataDrivenInvestorGym-DER: An environment for training RL agents to perform supervisory control of solar PV…There are hundreds of environments available for training Reinforcement learning agents (RL). However, only a limited number of them…Jul 8, 2019Jul 8, 2019
Siby Jose PlathottaminDataDrivenInvestorKaggle LANL earthquake challenge: Applying DNN, LSTM, and 1D-CNN Deep Learning modelsCan we try to solve the same problem with three different types of Deep neural network models? I was inspired by Siraj Raval’s excellent…Feb 25, 2019Feb 25, 2019
Siby Jose PlathottaminDataDrivenInvestorComparing GPU and TPU training performance on Google ColaboratoryI late September, Google made available their TPU as an accelerator option in Google Colaboratory. The example notebook provide…Nov 28, 20182Nov 28, 20182
Siby Jose PlathottaminDataDrivenInvestorExperimenting with a crazy new way to train Deep Neural Networks (part II)In my previous post, I presented some results from a training technique I developed. I compared mobilenet models (V1 and V2) trained on…Oct 10, 20181Oct 10, 20181
Siby Jose PlathottaminDataDrivenInvestorExperimenting with a crazy way to train Deep Neural NetworksThe history of Deep learning is full of bold and simple ideas. Sometimes they are genre-defining successes like GAN’s or fade out of…Oct 5, 2018Oct 5, 2018
