📣 The LLM shift no one saw coming!
Top engineers are dropping GPT for TinyLlama—faster, cheaper, and surprisingly more effective in real-world tasks 💼⚡
Want to know why?

👇 Read the article and future-proof your GenAI strategy:
👉 https://medium.com/@rogt.x1997/why-smart-engineers-are-ditching-gpt-for-tinyllama-and-you-should-too-345a042e6f6e

#TinyLlama #LLMTrends #GenAI2025 #OpenSourceAI
https://medium.com/@rogt.x1997/why-smart-engineers-are-ditching-gpt-for-tinyllama-and-you-should-too-345a042e6f6e

Just trained my own language model offline.
No cloud. No APIs. Fine-tuned it on my data, merged it, and ran it with llama.cpp.
This is what real AI literacy looks like.

Documentation:
https://github.com/hassanhabib/AI.Llama.Traing.Offline

Video:
https://youtube.com/watch?v=FQr7VrK5RRQ

#AI #LLM #LoRA #OfflineAI #TinyLlama

I asked #tinyllama to generate me a bio for my new #mastodon account at the great #OhaiSocial instance.

In short: it is a verbose inventor of text.

Me: "write me a bio info for my social media profile where my skills are presented: my skills are software, society, rocks , hiking,"

At least it stated: Here's an example of how you might incorporate inline citations into your bio information for your social media profile...

And I've forgot to say, the rest is also wrong.

#expanse #ai #experiment #knowledge #tinyllama #ollama

The excuse is, that this tinyllama is a VERY tiny model. llama3.2 works much smarter. (but harder to test because it already knows about the expanse without adding stuff to the "knowledge" so I need to invent something...)

Here it is how you can do #finetuning for a SMAL-language model that can be put on a #RaspberryPI or other edge-computing devices, or even wearables:
https://www.youtube.com/watch?v=DTYi7z4cLD0

#TinyLLaMA #TinyDolphin #Ollama #AIonEdge #MachineLearning #AIModels #EdgeComputing #AI #LLM

Impariamo ad eseguire i #LLM (#Large #Language #Model) localmente con #Ollama e entriamo nel mondo dell'#AI #generativa in maniera semplice e veloce:

https://www.risposteinformatiche.it/eseguire-ollama-e-vari-modelli-llm-su-linux/

#UnoLinux #Linux #gnu_linux #opensource #largelanguagemodel #phy3 #tinyllama #IntelligenzaArificiale #ChatGPT #ChatBot #Privacy

testing #Phi3, #TinyLlama and #Llama3 models on #raspberrypi using #ollama https://pimylifeup.com/raspberry-pi-ollama/

#LLM #SLM #AI #GenAI #Iot #EdgeAI

TinyLlama 1.1B: NEW LLAMA Model Size on 3 Trillion Tokens (Installation Tutorial)
Explore the future of language modeling with TinyLlama 🦙🌐! Unveiling a game-changing project with a colossal dataset of 3 trillion tokens, pushing AI boundaries! #TinyLlama #AIRevolution #NLP 🚀🤖
https://tweetclick.com/CnuQrhr5VM8

🌗 GitHub - jzhang38/TinyLlama
➤ TinyLlama 項目的特點和應用，以及其訓練詳細信息和速度比較。
✤ https://github.com/jzhang38/TinyLlama
本文介紹了 TinyLlama 項目，旨在在 3 兆令牌上預訓練 1.1B Llama 模型。TinyLlama 可以在許多建立在 Llama 上的開源項目中使用。此外，TinyLlama 具有緊湊性，僅有 1.1B 參數，可滿足對計算和內存佔用量有限的多種應用需求。
+ 這是一個非常有用的項目，尤其是對於那些需要緊湊且高效的語言模型的應用。
+ 這個項目的訓練速度非常快，而且可以在許多開源項目中使用，這是一個非常好的特點。
#GitHub #Llama #TinyLlama #pretraining #language models