Open models stripped bare by safety-busting tools – Fudzilla.com
…Weidmann added that he had removed safeguards from Google’s Gemma 4 model within 90 minutes of its release. Alice chief executive and co-founder Noam Schwartz said: “The genie is out…
Search
Related topics:
Google Gemma
All sources
xda-developers.com
29
developer.nvidia.com
15
blogs.nvidia.com
11
huggingface.co
9
androidauthority.com
6
screenrant.com
6
amd.com
5
wccftech.com
4
android-developers.googleblog.com
4
arstechnica.com
3
engadget.com
3
9to5mac.com
3
Tracked topic
Google Gemma
Gemma is a family of open-weight language models released by Google for text generation and related NLP tasks.
I tried Google's new DiffusionGemma, and watching it generate text like an image is unlike any local LLM
…DiffusionGemma is not a straight upgrade over Gemma 4 Google says quality is the trade-off The name makes DiffusionGemma sound like another member of the Gemma family, which it is, but…
I use my local LLM to triage my email every morning, but I'll still never let it send replies
…action is necessary. Gemma handles the busywork, and everything else that actually represents me remains entirely my responsibility. I wouldn't want it any other way. Ollama Released July 3, 2023 Developer…
I tested Google’s upcoming Gemini Nano 4 — its faster, smarter AI isn’t what I expected
…Still, the exciting news is that Gemma 4 has versions small enough to run on your smartphone. Specifically, Gemma 4 E2B and E4B are distilled down to effective two- and four-billion…
Discussions and forums
Gemma 4 MTP released
Blog post: https://blog.google/innovation-and-ai/technology/developers-tools/multi-token-prediction-gemma-4/ MTP draft models: https://huggingface.co/google/gemma-4-31B-it-assistant https://huggingface.co/google/gemma-4-…
google/gemma-4-12B · Hugging Face
Gemma is a family of open models built by Google DeepMind. Gemma 4 models are multimodal, handling text and image input (with audio supported on E2B, E4B, and 12B) and generating text output. This release includes open-w…
Show HN: Run Agent Skills with mistral.rs v0.8.10: /v1/skills support and more
Hey all! I'm the maintainer of mistral.rs. I just landed support for OpenAI-compatible Agent Skills via a /v1/skills endpoint, and it works with local open models.Until now Skills have basically been locked to closed mod…
20
2
The Qwen 3.6 35B A3B hype is real!!!
My personal test for small local LLM intelligence is to check whether a model has any ability to understand the code that I write for my own academic research. My research is on some pretty niche topics and I doubt that …
TextGen is now a native desktop app. Open-source alternative to LM Studio (formerly text-generation-webui).
Hi all, I have been making a lot of updates to my project, and I wanted to share them here. TextGen (previously text-generation-webui, also known as my username oobabooga or ooba) has been in development since December 2…
Top AI on Android updates for building intelligent experiences from Google I/O ‘26
…We’ve released a new test agent that allows you to experiment and debug your AppFunctions in a simulated agent environment. Early Access Program : Want to be among the first apps to…
I almost upgraded my GPU to run larger local LLMs, but this 8B model proved I didn't have to
…Qwen and Gemma have been my defaults Models are getting smaller and more efficient, but it also starts to feel confined A coworker recommended Qwen 3.5 9B and it was a…
AMD PACE - High-Performance Platform Aware Compute Engine
…Phi-4, OPT-6.7B, Llama-3.1-8B, Llama-3.2-3B, Qwen2-7B, Gemma-3-4B, and Gemma-3-12B . Speculative mode: AMD PACE speculative decoding achieves 2.5–3…
My 7-year-old GPU runs local AI perfectly, and I don't need my cloud subscriptions anymore
…The result is that my Turing-generation GPU, released well before ChatGPT existed, can now run models impressive enough to make a subscription to cloud providers entirely optional. If you're on…
Boost Inference Performance up to 15x on NVIDIA Blackwell Using DFlash Speculative Decoding | NVIDIA Technical Blog
…paper’s release, the research team has released 20 DFlash model checkpoints on Hugging Face with Blackwell and Hopper recipes, covering model families including Qwen, Kimi K2.6, Llama, Gemma, and gpt…
I replaced the expensive Claude Pro subscription with these local models, and my productivity didn’t drop a bit
…Gemma 4 E4B The ‘everyday’ champion If Qwen3.6 is my heavy-duty engineer, Gemma 4 E4B is my nimble, everyday champion. When it comes to local LLMs, we often think that…