Omar Sanseviero (@osanseviero)

Gemma 4 출시를 위해 Hugging Face, vLLM, llama.cpp, Ollama, NVIDIA, Unsloth, Cactus, SGLang, Docker, Cloudflare 등 여러 AI 생태계 파트너와 협업했다는 내용이다. 최신 모델 배포에 필요한 오픈소스·인프라 협력이 강조된 중요한 발표다.

https://x.com/osanseviero/status/2041154555530932578

#gemma #opensource #llm #vllm #ollama

Omar Sanseviero (@osanseviero) on X

People underestimate the level of collaboration that needs to happen for a model such as Gemma 4 to land Before the launch, we worked with HF, VLLM, llama.cpp, Ollama, NVIDIA, Unsloth, Cactus, SGLang, Docker, CloudFlare, and so many others This ecosystem is amazing 🔥

X (formerly Twitter)

AshutoshShrivastava (@ai_for_success)

Pixel 10 Pro에서 Gemma 4 E2B를 사용해 오디오 전사를 수행한 사례를 공유한다. 현재는 최대 30초까지 지원하며, 온디바이스 AI 활용의 실사용 가능성을 보여주는 흥미로운 적용 사례다.

https://x.com/ai_for_success/status/2041177663029784704

#gemma #transcription #pixel10pro #ondeviceai #audio

AshutoshShrivastava (@ai_for_success) on X

Using Gemma 4 E2B for audio transcription on my Pixel 10 Pro. It support max 30 sec for now.

X (formerly Twitter)

merve (@mervenoyann)

Gemma 4가 추론을 위한 thought channel을 제공한다는 팁입니다. 추론과 답변 데이터를 활용해 fine-tuning하면, 생각 과정을 주입하는 방식으로 성능을 개선할 수 있다고 설명합니다.

https://x.com/mervenoyann/status/2041183790375002259

#gemma #reasoning #finetuning #llm #opensource

merve (@mervenoyann) on X

tip: Gemma 4 exposes a thought channel for reasoning if you fine-tune it on reasoning and answers, you can inject thinking channel like below ⤵️

X (formerly Twitter)

clem (@ClementDelangue)

Gemma 4가 Hugging Face 랭킹 1위를 차지했다는 내용으로, 최신 오픈 모델의 성능과 인기를 보여주는 주목할 만한 소식이다.

https://x.com/ClementDelangue/status/2040911131108069692

#gemma #huggingface #llm #opensource #aimodel

clem 🤗 (@ClementDelangue) on X

Gemma 4 is #1 on @huggingface!

X (formerly Twitter)
🌖 GitHub - kessler/gemma-gem:透過 WebGPU 在瀏覽器端實現 Google Gemma 4 模型運行,無需雲端 API 或數據傳輸
➤ 極致隱私的網頁 AI 助手:在瀏覽器中完全本地化運行 Gemma 4
https://github.com/kessler/gemma-gem
Gemma Gem 是一款創新的瀏覽器擴充功能,旨在將 Google 的 Gemma 4 模型直接部署於用戶終端。該工具完全依賴 WebGPU 技術,無需依賴雲端伺服器或第三方 API,確保所有數據均保留在本地設備上。使用者透過此擴充功能,不僅能與網頁內容進行對話,還能執行讀取頁面、自動點擊按鈕、填寫表單及執行 JavaScript 等操作。其架構巧妙結合了 Offscreen Document 與 Content Script,在保障隱私的同時,為網頁互動提供了高效的 AI 代理服務。
+ 太驚人了!終於能不用 API key 就在瀏覽器裡運行模型了,雖然下載模型要點空間,但為了隱私這很
#Chrome 擴充功能 #AI #WebGPU #Gemma 4 #瀏覽器自動化
GitHub - kessler/gemma-gem: Gemma Gem runs Google's Gemma 4 model entirely on-device via WebGPU — no API keys, no cloud, no data leaving your machine.

Gemma Gem runs Google's Gemma 4 model entirely on-device via WebGPU — no API keys, no cloud, no data leaving your machine. - kessler/gemma-gem

GitHub

Now you can run AI models like Gemma 4 Directly in your phones easily with Google's Open Source App Google AI Edge Gallery , It supports multiple AI models you download in the app itself from Hugging face and it work fully Offline.

https://firethering.com/google-ai-edge-gallery-offline-llm-app/

#opensource #google #gemma #ai #trending #lllm

Google AI Edge Gallery: Run LLMs Offline on Your Phone

Google AI Edge Gallery lets you run open-source LLMs straight on your phone. No cloud. Once you download the models, you're offline. You get chat, image analysis, audio transcription, prompt testing. All on-device. Newer models like Gemma 4 mean better reasoning and multimodal stuff on mobile hardware. It’s more like a sandbox where you can test, run, and compare models directly on your device.

Firethering

yeah I got an opinion on this week's #GEMMA 4 from Google -- and forced update to LM-Studio-0.4.8+ to use it:

https://installers.lmstudio.ai/win32/x64/0.3.36-1/LM-Studio-0.3.36-1-x64.exe

__installers.lmstudio.ai/win32/x64/0.3.36-1/LM-Studio-0.3.36-1-x64.exe

* good luck downloading it otherwise

** URL tampering (without path traversal) is your friend

*** thank you Linux, for showing me the way:

__installers.lmstudio.ai/linux/x64/0.3.36-1/LM-Studio-0.3.36-1-x64.AppImage #Gemma4

Prince Canuma (@Prince_Canuma)

MLX 기반 멀티모달 비전-언어 도구 mlx-vlm v0.4.4에 TurboQuant 성능 개선이 대폭 적용됐고, Open Evals 벤치마크에서 Gemma 4 26B IT를 M3 Ultra로 테스트한 결과 품질 저하 없이 동일한 78% 정확도를 기록했다고 소개한다.

https://x.com/Prince_Canuma/status/2040877782922649865

#mlx #benchmark #quantization #gemma #multimodal

Prince Canuma (@Prince_Canuma) on X

TurboQuant: Open Evals on MLX 🔥 Yesterday I launched mlx-vlm v0.4.4 with major TurboQuant performance improvements. Today, the open benchmark results on MM-NIAH (val, 520 samples) using Gemma 4 26B IT by @GoogleDeepMind on M3 Ultra: → 0 quality loss — 78% accuracy for both

X (formerly Twitter)

田中義弘 | taziku CEO / AI × Creative (@taziku_co)

Google의 Gemma 4가 iPhone 17 Pro에서 완전 온디바이스로 동작하며, 이미지 이해와 추론, 필요 시 E2B 연동까지 수행한다고 소개됐다. MLX 최적화로 약 40 tok/s 성능을 내며, 네트워크나 클라우드 없이 실행되는 점이 핵심이다.

https://x.com/taziku_co/status/2040934086681006206

#google #gemma #ondevice #llm #mlx

田中義弘 | taziku CEO / AI × Creative (@taziku_co) on X

GoogleのGemma 4、 iPhone 17 Proで完全オンデバイス動作。 画像を見て、推論して、必要ならE2Bまで回す。 しかもMLX最適化で約40tok/s。 ネット不要、クラウド不要、機内モードがデフォルト。

X (formerly Twitter)