Reddit Tech VN BotKhám phá mô hình AI phi2 của Microsoft, phù hợp để chạy trên PC với 12GB RAM + 3GB VRAM + GTX 1050 + Linux Mint. Phi2 được lượng tử hóa Q4K, tối ưu hiệu suất trên GPU trung bình. Thử tải về từ Hugging Face hoặc TheBloke và trải nghiệm mô hình AI phi-commercial này! #AIModel #Linux #TechVietnam #LocalLLaMA #Phi2 #GPUOptimization #AICommunity
Mô hình AI 2,7B (Phi-2) tạo ra câu trả lời đáng kinh ngạc, vượt qua khả năng dự đoán. Nó giữ được giọng điệu, tạo ra lý lẽ đa đoạn, tránh ảo giác và giữ vững căn cứ. #AI #TríTuệNhânTạo #MôHìnhPhi2 #Phi2 #LocalLLaMA #MachineLearning #HọcMáy #TríTuệNhânTạoViệtNam
https://www.reddit.com/r/LocalLLaMA/comments/1ozpq3u/this_response_is_from_a_27b_model_phi2_i_dont/
Dirk WoutersWer hat an der Uhr gedreht…?
Jede CPU braucht einen Taktgeber, um arbeiten zu können. Der Commodore C64 hat jedoch mehr als einen davon. Ich versuche, die verschiedenen Taktgeber in diesem Artikel zu erklären.
#c64 #Clock #Clocks #Color #commodore #Expansion #Expansionport #NTSC #PAL #Phase #Phi0 #Phi2
framboise314Et si on installait une #IA en local sur un @raspberrypi RASPBERRYPI5 ? Je vous propose d'installer llama-cpp et #Ollama sur la framboise et de la tester avec différents modèles comme #Phi2 ou #Mistral
https://www.framboise314.fr/faites-tourner-un-llm-de-type-chatgpt-sur-le-raspberry-pi-5/
https://www.framboise314.fr/faites-tourner-un-llm-de-type-chatgpt-sur-le-raspberry-pi-5/
Faites tourner un LLM de type ChatGPT sur le Raspberry Pi 5 - Framboise 314, le Raspberry Pi à la sauce française....
Dans de précédents articles je vous ai expliqué comment mettre en œuvre l'IA sur le Raspberry Pi 5 : Module Hailo-8L, module Coral TPU, TensorFlow...
brozu ▪️🧠 #NVIDIA researchers are advancing SLMs through structured weight pruning and knowledge distillation, cutting down model size while preserving performance.
For instance, #Minitron 8B and 4B models, derived from Nemotron 15B, outperform many models trained from scratch. Despite its size, it competes with top-tier models like #Gemma2 and #Phi2.
How to Prune and Distill Llama-3.1 8B to an NVIDIA Llama-3.1-Minitron 4B Model | NVIDIA Technical Blog
Large language models (LLM) are now a dominant force in natural language processing and understanding, thanks to their effectiveness and versatility. LLMs such as Llama 3.1 405B and NVIDIA Nemotron-4…
5h15hnow trying out Microsoft #Phi2 with Ollama https://jussiroine.com/2024/01/getting-started-with-ollama-with-microsofts-phi-2/
#GenAI #LLM #SLM #AI #GenerativeAI
https://jussiroine.com/2024/01/getting-started-with-ollama-with-microsofts-phi-2/
Getting started with Ollama with Microsoft's Phi-2
It seems that each week brings a dozen new generative AI-based tools and services. Many are wrappers to ChatGPT (or the underlying LLMs such as GPT 3.5 Turbo), while some bring much more. Ollama is one of the latter, and it's amazing. In this blog post, I'll briefly examine
Preston von GabbleduckTell me your #LLM has been trained using cishet data, not vetted for diversity & inclusion, without telling me your LLM has been trained using cishet data & vetted for diversity & inclusion. #Phi #Phi2
Erik JonkerImpressed with the ease of use of https://lmstudio.ai/
, easy to install, after that you can download a version of the Phi-2 model to play around with. All locally without extra costs or internet connection.
#LLM #generativeAI #Phi2 #experiment #microsoft #LMstudio #AI
, easy to install, after that you can download a version of the Phi-2 model to play around with. All locally without extra costs or internet connection.
#LLM #generativeAI #Phi2 #experiment #microsoft #LMstudio #AI
Gea-Suan Lin微軟 Phi-2 model 的授權改成 MIT License
在 Hacker News 的「Microsoft Phi-2 model changes licence to MIT (huggingface.co)」這邊看到的消息，連結是改成 MIT License 的 commit：「Upload 3 files · microsoft/phi-2 at 7e10f3e」。
看了一下 mo
#Computer #Murmuring #Network #ai #language #large #learning #license #llm #machine #microsoft #mit #model #network #neural #phi2
Alessio Pomaro🧠 #Microsoft ha rilasciato #Phi2, un #LLM di "piccole" dimensioni, ma dalle performance elevate.
💡 La direzione è proprio questa.. innovazione nel ridimensionamento dei modelli e nella pulizia dei dati di training.
🦾 Le prestazioni si avvicinano a quelle di modelli molto più grandi, come #Llama2, Mistral e Gemini Nano.
#AI #GenAI #IntelligenzaArtificiale
https://www.microsoft.com/en-us/research/blog/phi-2-the-surprising-power-of-small-language-models/
Phi-2: The surprising power of small language models
Phi-2 is now accessible on the Azure model catalog. Its compact size and new innovations in model scaling and training data curation make it ideal for exploration around mechanistic interpretability, safety improvements, and fine-tuning experimentation on a variety of tasks.