[18:35] Dit zijn Gemma en Mateusz, de nieuwe eigenaren van vakantiepark Het Wylde Pad in Twijzelerheide

Het veelbesproken vakantiepark Het Wylde Pad in Twijzelerheide heeft nieuwe eigenaren: Gemma (33) en Mateusz (45) Sęk-van Norden. Wat zijn ze ermee van plan?

https://lc.nl/friesland/achtkarspelen/Dit-zijn-Gemma-en-Mateusz-de-nieuwe-eigenaren-van-vakantiepark-Het-Wylde-Pad-46142902.html

#HetWyldePad #Twijzelerheide #Gemma #33 #Mateusz #45 #Norden

DolphinGemma: Google AI model understands dolphin chatter https://www.artificialintelligence-news.com/news/dolphingemma-google-ai-model-understands-dolphin-chatter/ #google #pixel #gemma #mobile #ai #tech #news #technology

@theaigrid

#google #ai #gemma #gemini #lmm

Google Just 'Took Over' (next week someone else ) #AI....(Wow)"

https://youtu.be/MqmVmZ1HbKk?feature=shared

Zum heutigen #Tassenfreitag mache ich noch einen großen KI Vergleich. Ich lasse folgende KIs eine Bildbeschreibung einer Tasse mit etwas schwieriger Schrift drauf beschreiben. Es gehen ins Rennen:

#ChatGPT
Le Chat von #Mistral
#DeepSeek
#Pilama mit #Gemma 3.4

Wer macht es am Besten?

Whoohoo! #LMStudio lets me run the big Gemma 3 27b, 4-bit quantized locally on my slightly old gaming laptop with Geforce 3080 TI. It is slow, but my first tests show it is indeed fairly powerful. Not up to the reasoning models for reasoning tasks (duh!), but way above everything else I could run locally before (Llama 3).

A big step towards keeping your data private and still enjoy the services of a great LLM.

Introducing Gemma 3: The Developer Guide.

https://developers.googleblog.com/en/introducing-gemma3/

#Gemma 3 Released with #Multilingual & #Multimodal Capabilities

Blog: https://developers.googleblog.com/en/introducing-gemma3/
Video: https://youtu.be/UU13FN2Xpyw

#ai #llm #Google

Gemma 3: Google launches its latest open AI models https://www.artificialintelligence-news.com/news/gemma-3-google-launches-its-latest-open-ai-models/ #google #opensource #developers #gemma #ai #tech #news #technology

https://www.europesays.com/1775007/ What’s Irish about Italy? More than you think! #Antica #Daytrip #Donato #festival #firenze #Florence #Gemma #Hotel #ireland #Irish #italy #jazz #LaGemma #Lucca #Pompeii #san #Torre

Check this screenshot and let it sink in:

These are the real probabilities of the next token prediction from #Gemma #LLM for the prompt "The reliability of Wikipedia is very..." (from this video: https://www.youtube.com/watch?v=UGO_Ehywuxc)

Look at the difference between next word "high" and "low". Or between "poor" and "good". You have about 1% split chance between reading two very opposite statements. I think it nicely shows that we should never take these #AI outputs as "answers", they have no basis in facts.

Corriere.it - Homepage by di Federica Bandirali
Tina Cipollari oggi a Belve: l'esordio a Uomini e donne, la separazione da Chicco e gli scontri con Gemma

Translated:
Tina Cipollari's appearance today on Belve: her debut on Men and Women, her separation from Chicco, and her clashes with Gemma.

#TinaCipollari #Belve #MenandWomen #Italy #Gemma #Chicco
https://www.corriere.it/spettacoli/24_dicembre_10/tina-cipollari-belve-c918123c-41d9-4f75-b1ac-3b8439476xlk.shtml

https://youtu.be/rXNBFTT1pHk
#monsterhunterwilds #mh #mhw #monsterhunter #capcom #gemma

https://youtu.be/UUcyxe_bbuQ

Can Large Language Models (LLMs) truly reason? No.

"We found no evidence of formal reasoning in language models including open-source models like #Llama, #Phi, #Gemma, and #Mistral and leading closed models, including the recent #OpenAI #GPT-4o and #o1-series.

Their behavior is better explained by sophisticated pattern matching—so fragile, in fact, that changing names can alter results by ~10%!

We can scale data, parameters, and compute—or use better training data for Phi-4, Llama-4, GPT-5.

But we believe this will result in 'better pattern-matchers,' not necessarily 'better reasoners."

Full paper https://arxiv.org/pdf/2410.05229

#AI #LLM #Ollama #llama3 #Gemma #jan
我拒絕毫無透明性&封閉的ChatGPT,因為這種線上AI無自訂性還有言論審查限制使用者自由
封閉性也阻止使用者檢查AI模型的數據集或技術運作是否侵犯版權和隱私

https://jan.ai https://ollama.com
在自己的Linux電腦用自由開源軟體Jan/Ollama離線執行開源大型語言模型AI
https://codeberg.org/wikifox/pages
這種方式的好處是可自訂性更高,還能下載無言論審查的版本(例如llama2 uncensored)
Llama 3.1 8B和Gemma2 9B以及llama 3 taiwan 8b都滿好用的

問好幾個模型把它們的回答都列入考慮,就像請好幾個工讀生幫忙解決生活問題一樣,長期使用省下來的時間精力價值會超過我買電腦的花費

Poco a poco los modelos de lenguaje abierto y con modo de ejecución local (en nuestras máquinas, y no dependiendo de la "nube") poseen un mayor impacto en la comunidad, no solo por un tema de costo, también por un tema de privacidad y personalización.

#llama #mistral #phi #gemma #llm #ia
Chao #chatgpt
@hn50 https://social.lansky.name/@hn50/113176019392578208

Google’s tiny AI model ‘Gemma 2 2B’ challenges tech giants in surprising upset https://venturebeat.com/ai/googles-tiny-ai-model-gemma-2-2b-challenges-tech-giants-in-surprising-upset/ #AI #Gemma

Been playing around with MediaPipe and Gemma-2b in an iOS app built using SwiftUI and Combine.

Interestingly the results I get from the same prompts used in the Android app give different and more tempermental results.

You can check out the code here: https://github.com/DarrylBayliss/Simon-Says-iOS