Robert Dadashi (@robdadashi) / X

Robert Dadashi

220 posts

Robert Dadashi

@robdadashi

Gemma TL @GoogleDeepMind

Paris, France

ddsh.github.io

Joined September 2014

Pinned
Robert Dadashi
@robdadashi
Apr 3
Gemma 4 is out! 🔥 Once again, our team has delivered best-in-class open weights models, fully available for research and commercial use!
Demis Hassabis
@demishassabis
Apr 2
Excited to launch Gemma 4: the best open models in the world for their respective sizes. Available in 4 sizes that can be fine-tuned for your specific task: 31B dense for great raw performance, 26B MoE for low latency, and effective 2B & 4B for edge device use - happy building!
Robert Dadashi
@robdadashi
Apr 8, 2024
I am very happy to announce that Gemma 1.1 Instruct 2B and “7B” are out! Here are a few details about the new models: 1/11
Robert Dadashi
@robdadashi
May 2, 2019
Just released a notebook to generate the figures in "The Value Function Polytope in RL": bit.ly/2ZRmurc. It's fun to play with !
Robert Dadashi
@robdadashi
Jun 27, 2024
I am so proud to announce that: - Gemma 2 27B IT tops all open weights models on Chatbot Arena, with a pinch of optimism in the face of uncertainty :) - Gemma 2 9B IT sets a new frontier for models of similar size. 1/n
Robert Dadashi
@robdadashi
Dec 15, 2021
Very proud of our latest work: AQuaDem - Action Quantization from Demonstrations. The idea is simple: 1- Learn a state-conditioned quantization of a continuous action space from human demonstrations 2- Learn a controller in the induced MDP with a discrete action method, e.g DQN
00:00
Robert Dadashi
@robdadashi
Jun 9, 2020
New paper out: PWIL ! A simple imitation learning method, which reinforces a reward signal based on a distance to expert demonstrations. Makes Humanoid walk with a single demonstration (below). 1/ arxiv.org/abs/2006.04678
00:00
Robert Dadashi
@robdadashi
Jul 2, 2020
We have just released the code for our latest paper on Imitation Learning: PWIL (arxiv.org/abs/2006.04678). It’s simple and concise and yet performs strongly with MuJoCo environments. The code builds on @deepmind’s Acme (which is great !) github.com/google-researc…
Robert Dadashi
@robdadashi
Apr 22, 2019
During my @GoogleAI residency I have been fortunate to work with my research mentor @marcgbellemare and other great collaborators on projects that I am weirdly excited about. This led to 2 papers accepted at ICML: (1/2)
Robert Dadashi
@robdadashi
Feb 21, 2024
I am so proud to see Gemma released today! I have had a fantastic time working on post-training and RLHF with an amazing team. Cannot wait to see what the community builds with these models!
Google DeepMind
@GoogleDeepMind
Feb 21, 2024
Introducing Gemma: a family of lightweight, state-of-the-art open models for developers and researchers to build with AI. 🌐 We’re also releasing tools to support innovation and collaboration - as well as to guide responsible use. Get started now. → dpmd.ai/3UJu1Y1
Robert Dadashi
@robdadashi
Apr 8, 2024
Replying to @robdadashi
The training data was pretty much the same as v1.0, but we switched the RL algorithm to something new. I hope that we will be able to disclose more about this in the future :). 6/11
Robert Dadashi
@robdadashi
Jun 11, 2019
I will talk about the *mysterious* polytopes in reinforcement learning at #ICML2019, Tuesday June 11th at 5:15pm in room 104 and at 6:30 at poster 119.
Robert Dadashi
@robdadashi
Sep 23, 2022
Very proud to contribute to making RL agents more accessible and reproducible!
Google DeepMind
@GoogleDeepMind
Sep 23, 2022
Acme, a framework for distributed RL research, has been updated to be cleaner, more modular, and to support more agents - including offline & imitation. Try it yourself! GitHub: dpmd.ai/acme-github Quickstart: dpmd.ai/acme-quickstart V2 Paper: dpmd.ai/acme-paper 1/
Robert Dadashi
@robdadashi
Apr 8, 2024
Replying to @robdadashi
Similarly to v1.0, we enforced verbosity penalty on the models at training time even though it means worse performance on benchmarks. If you still feel like Gemma models are too chatty, prompting with a target word count can help. 7/11
Robert Dadashi
@robdadashi
Feb 25, 2019
Proud to be part of this project tackling how we should think of statistics propagation & derive new algorithms in the context of distributional reinforcement learning
Marc G. Bellemare
@marcgbellemare
Feb 25, 2019
Mark Rowland's distributional RL paper on samples and statistics (& potential mismatch) is out -- big step towards understanding the method w/ @wwdabney @RobertDadashi S. Kumar R. Munos arxiv.org/abs/1902.08102