Sven Mika
Reinforcement Learning (RLlib) and LLMs
This profile was put together from Sven's public GitHub portfolio, and they haven't joined Vibetown yet, so reach them through their portfolio. Is this you? Contact us to update or remove it.
About
Sven Mika works in ML & data science and AI engineering, based in San Francisco, CA; Düsseldorf, Germany (anyscale.com). Reinforcement Learning (RLlib) and LLMs. Public projects on GitHub include Transformer Rlaif End To End (Pre-train a small transformer with ray on your laptop from scratch on a translation task, then post-train…), Dreamer V3 (Implementation (TensorFlow/keras) of the DreamerV3 model-based RL algorithm by Hafner et al. 2023) and Shine ([s]erver-[h]osted [i]ntelligent [n]eural-net [e]nvironment), mostly written in Python, Jupyter and C++.
Projects 6
Transformer Rlaif End To End
Pre-train a small transformer with ray on your laptop from scratch on a translation task, then post-train with RLAIF.
github.comDreamer V3
Implementation (TensorFlow/keras) of the DreamerV3 model-based RL algorithm by Hafner et al. 2023
github.comShine
[s]erver-[h]osted [i]ntelligent [n]eural-net [e]nvironment
github.comAttention Is All You Need
Implementation of the Transformer Model described in "Attention is all you need" by Vaswani et al.
github.comPerceptron To Dnc
"From Perceptrons to Differentiable Neural Computers", by Dr. Sven Mika. Presentation at DUS ML meetup Dec 2nd 2017
github.comRLlib_UE5_Demo
RLlib + Unreal Engine 5 demo
github.com