Svilen Todorov
Blog Fiction About

In the Media

An interview I gave for The Register:

Could OpenAI's 'too dangerous to release' language model be used to mimic you online?


Posts

  • Mar 22, 2023
    Simple Three Factors Framework for Goal Achievement
  • Mar 20, 2023
    Exploring Lichess data on improvement using e4 vs d4 and blitz vs rapid
  • Jul 21, 2021
    As Many Times As It Takes
  • Jul 8, 2020
    The Unwatchability of Yesterday's Videos Today and the Promise of Tomorrow
  • Mar 27, 2020
    StyleGAN for Evil: Trypophobia and Clockwork Oranging
  • Mar 18, 2020
    Training Google's Reformer - takeaways, code, and weights
  • Mar 12, 2020
    Corona Advice for People Who Go Against Advice
  • Feb 8, 2020
    Tips and Pain Point Workarounds for Machine Learning Work in Google Colab
  • Jan 23, 2020
    Talking to Myself or How I Trained GPT2-1.5b for Rubber Ducking using My Facebook Chat Data
  • Jan 19, 2020
    How is search so goddamn bad? A ‘Case Study’
  • May 8, 2019
    Fine-tunning OpenAI's (Larger) GPT-2-345M on Conversation data (Update)
  • Apr 29, 2019
    A Proposed Task for Training a Human-Imitating System as a Benchmark and Potential Part of The Road to AGI
  • Mar 22, 2019
    Adding Layers to the middle of a pre-trained network whithout invalidating the weights
  • Mar 12, 2019
    Generating Fake Conversations by fine-tunning OpenAI's GPT-2 on data from Facebook Messenger
  • Jul 31, 2018
    Building an in-browser Rock Paper Scissors Neural Network capable of beating humans

subscribe via RSS


Demo

Demo - Building an in-browser Rock Paper Scissors Neural Network capable of beating humans

  • Svilen Todorov
  • sviltodorov@gmail.com
  • tenoke

Personal site for Data Scientist and Machine Learning Engineer - Svilen Todorov