Jeremiah Harmsen (@jeremiahharmsen) 's Twitter Profile
Jeremiah Harmsen

@jeremiahharmsen

Creator of #TensorFlowHub and @TensorFlow Serving.

Lead in Google Brain.

ID: 923553255652843521

calendar_today26-10-2017 14:13:49

872 Tweet

1,1K Followers

523 Following

Google AI Developers (@googleaidevs) 's Twitter Profile Photo

PaliGemma 2 mix is an upgraded vision-language model that supports image captioning, OCR, image Q&A, object detection, and segmentation. With sizes from 3B-28B parameters, there's a model for everyone. Get started. → goo.gle/430HnDe

PaliGemma 2 mix is an upgraded vision-language model that supports image captioning, OCR, image Q&A, object detection, and segmentation. With sizes from 3B-28B parameters, there's a model for everyone. Get started. → goo.gle/430HnDe
Andreas Steiner (@andreaspsteiner) 's Twitter Profile Photo

Looking for a small or medium sized VLM? PaliGemma 2 spans more than 150x of compute! Not sure yet if you want to invest the time 🪄finetuning🪄 on your data? Give it a try with our ready-to-use "mix" checkpoints: 🤗 huggingface.co/blog/paligemma… 🎤 developers.googleblog.com/en/introducing…

Looking for a small or medium sized VLM? PaliGemma 2 spans more than 150x of compute!

Not sure yet if you want to invest the time 🪄finetuning🪄 on your data? Give it a try with our ready-to-use "mix" checkpoints:

🤗 huggingface.co/blog/paligemma…
🎤 developers.googleblog.com/en/introducing…
Michael Tschannen (@mtschannen) 's Twitter Profile Photo

📢2⃣ Yesterday we released SigLIP 2! TL;DR: Improved high-level semantics, localization, dense features, and multilingual capabilities via drop-in replacement for v1. Bonus: Variants supporting native aspect and variable sequence length. A thread with interesting resources👇

📢2⃣ Yesterday we released SigLIP 2! 

TL;DR: Improved high-level semantics, localization, dense features, and multilingual capabilities via drop-in replacement for v1.

Bonus: Variants supporting native aspect and variable sequence length.

A thread with interesting resources👇
Clément (@clmt) 's Twitter Profile Photo

Gemma 3 is out! We are focused on bringing you open models with best capabilities while being fast and easy to deploy: - 27B lands an ELO of 1338, all the while still fitting on 1 single H100! - vision support to process mixed image/video/text content - extended context window

Gemma 3 is out! 

We are focused on bringing you open models with best capabilities while being fast and easy to deploy:

- 27B lands an ELO of 1338, all the while still fitting on 1 single H100!
- vision support to process mixed image/video/text content
- extended context window
Omar Sanseviero (@osanseviero) 's Twitter Profile Photo

Introducing: ShieldGemma 2 - a 4B model for image safety classification 🛡️ 👀Use as input filter for VLMs ❌or for blocking dangerous image generation outputs 🦺Great for production deployments Blog: developers.googleblog.com/en/safer-and-m… Model: huggingface.co/google/shieldg…

Introducing: ShieldGemma 2 - a 4B model for image safety classification 🛡️

👀Use as input filter for VLMs
❌or for blocking dangerous image generation outputs
🦺Great for production deployments

Blog: developers.googleblog.com/en/safer-and-m…
Model: huggingface.co/google/shieldg…
Peter Battaglia (@peterwbattaglia) 's Twitter Profile Photo

We're hiring a Research Engineer in AI for Sustainability Google DeepMind (San Francisco / Mountain View). Seeking strong engineers at the interface of machine learning, environmental sustainability, weather, dynamical systems, and/or remote sensing: boards.greenhouse.io/deepmind/jobs/…

Alexandre Ramé (@ramealexandre) 's Twitter Profile Photo

Hiring two student researchers for Gemma post-training team at Google DeepMind Paris! First topic is about diversity in RL for LLMs (merging, generalization, exploration & creativity), second is about distillation (with Nino Vieillard). Ideal if you're finishing PhD. DMs open!

meg.ai 🇨🇦 (@meganrisdal) 's Twitter Profile Photo

It's interesting to see PaliGemma 2 & Gemma 2 among the top most popular models used to build Kaggle Package solutions for the SVG image generation competition Why not much Gemma 3? kaggle.com/competitions/d…

It's interesting to see PaliGemma 2 &amp; Gemma 2 among the top most popular models used to build <a href="/kaggle/">Kaggle</a> Package solutions for the SVG image generation competition

Why not much Gemma 3? kaggle.com/competitions/d…
Google DeepMind (@googledeepmind) 's Twitter Profile Photo

Our Senior Research Director Joelle Barral joins @fryrsquared to talk about how AI is being developed to improve healthcare. 🏥 From increasing access to life-saving screening tools, to more personalized treatments, we’re excited for AI’s potential to support patient

Deep Learning Indaba (@deepindaba) 's Twitter Profile Photo

Last call to submit a paper at DLI 2025 ! ⌛ Accepted papers will be presented during the Research in Africa Showcase sessions, providing a platform for engagement with leading AI scholars and industry experts. 🔥 Apply before May 20, 2025, at 11:59 PM AoE

Yi Xu (@_yixu) 's Twitter Profile Photo

🚀Let’s Think Only with Images. No language and No verbal thought.🤔 Let’s think through a sequence of images💭, like how humans picture steps in their minds🎨. We propose Visual Planning, a novel reasoning paradigm that enables models to reason purely through images.

🚀Let’s Think Only with Images.

No language and No verbal thought.🤔 

Let’s think through a sequence of images💭, like how humans picture steps in their minds🎨. 

We propose Visual Planning, a novel reasoning paradigm that enables models to reason purely through images.
👩‍💻 Paige Bailey (@dynamicwebpaige) 's Twitter Profile Photo

✨🎶 Massive potential for Google DeepMind's Music AI Sandbox, so glad that these products are being developed in collaboration with artists and musicians! #GoogleIO deepmind.google/discover/blog/…

Google AI Developers (@googleaidevs) 's Twitter Profile Photo

2️⃣SignGemma is a sign language understanding model that’s coming later this year 🤟🏼It’s a massively multilingual model that’s best at translating ASL into English text, enabling further development of tech access for Deaf and Hard of Hearing users. 🧏 Share your feedback and

Demis Hassabis (@demishassabis) 's Twitter Profile Photo

Amidst the massive demand for Gemini 2.5 and Veo 3 models, wanted to also give a big shout out to our world-class infrastructure, chip and SRE teams, who work tirelessly to keep our wonderful TPUs from melting, and without whose incredible work none of this would be possible.