#phi3 hasil pencarian
Here is a simple demo where I run #Phi3 locally on my phone completely disconnected. The ability to run #SLMs on the edge opens up new opportunities. I show a few examples of different generations and achieve ~8-10 TPS. This is using a qunatized 128K version of the model. 🤟
Microsoft is redefining what's possible with SLMs—#phi3. Phi-3 models are privacy-focused but can handle only simple tasks. Yet, at times, they significantly outperform language models of the same and larger sizes on key benchmarks.
Exciting Update! Microsoft introduces powerful new models to the Phi-3 family on Azure. Discover more: tinyurl.com/h7nuhcpy #Phi3 #Azure #Phi3
Here's a nice GUI to test the new models #Phi3 #Llama3 #Mixtral #OpenSource #LLM yourself: @LMStudioAI. Here's how 🧵
Let’s cook with Quantized Phi models using MLX. Phi-3-mini-4k-instruct-8bit French omelette 🍳 Prompt: 203.066 tps Generation: 70.524 tps M3 Max: 64 GB. #mlx #phi3 @awnihannun
🥳Multi-modal Phi-3-mini is here! #LLaVA-Phi-3-mini outperforms LLaVA-v1.5-7B and matches the performance of LLaVA-Llama-3-8B in multiple benchmarks. 😊For easy applications, #GGUF weights are provided. 👉github.com/InternLM/xtuner @_akhaliq #Phi3
Did you know the Phi-3-mini-4k-instruct-onnx model can handle prompts up to 4k words? @Lee_Stott explains how using NLP techniques, from setting up Python to utilizing the generate() API, providing step-by-step instructions and code examples. buff.ly/3K3JFGQ #phi3 #AI
Phi-3 failed at my first test: Number of occurrences and positions of the letter "r" in "blueberry" - usually I can get an LLM to finally accept that there are Rs in blueberry but after 20 minutes, I gave up and told it there were two. Apparently, we had a misunderstanding. #Phi3
Attention $TAO holders, this is big!🤖 The critical energy problem Mark mentioned is now being solved by SN29 @Coldint_ Large language models once needed MASSIVE compute, but @Coldint_ improved #LLaMA & #Phi3, letting us scale models—without increasing loss Let me explain🧵👇
#PHI3 'Microsoft launches PHI-3: What are SLMs, and what can they do?' :Explained by Mr Bijin Jose @bijin_jose - SLM Small Language Model #Microsoft #SLMs #LLMs #LLM #ArtificialIntelligence #GenAI #Parameters #token #Specialisation #Latency #technology #UPSC Source: IE
Check out Microsoft's latest Phi-3 models, starting with Phi-3-mini. They're versatile, running on various devices, and support different model formats. Learn more here: techcommunity.microsoft.com/t5/microsoft-d… #phi3 #AI
Microsoft Phi3 is coming, if you want to inference Phi3-mini model , please read my blog aka.ms/phi3gettingsta… 🥳🥳🥳🥳🥳 #phi3 #SLM #wasmedge @ollama @secondstateinc
Trying out #phi3 (medium). It's mighty fast! Almost got the correct answer (lol include string.ptr) but it could make a nice model for copilot-ing, at 14b...
Can you match the #LLM to its response? 😅 Four leading LLMs (#llama3, #phi3, #gemma, and #mistral) were given the same prompt, and below are the results.
🖼️Vision Model on #LLM that can run in your cell-phone and works like a champ! awesome! 👏@intern_lm for producing this #LLaVA model fine-tuned from #Phi3 by @MSFTResearch #ai4all #ai4good
Video summaries just got easier! ✨ Azure AI Video Indexer & Phi-3 deliver concise, local insights while keeping your data secure. Learn more: msft.it/6011lRY5d #Phi3 #VideoSummarization #AzureAI #AzureVideoIndexer
Why is AI so bad when it comes to dates? I tried this with GPT 3.5 a while ago and had the same issue. This now is with Phi-3 Mini They all have completely incorrect calendar knowledge (April 1, 2024 was a MONDAY) #ai #chatgpt #phi3
" Meta's #Llama3, Microsoft's #Phi3 and Hugging Face's #Zephyr when fine-tuned on specific tasks outperform OpenAI's #GPT4! 🤯" linkedin.com/posts/pieromol…
🌻おはようございます！ 連休2日目、創造性が高まる日になりそうです🎨 【今日のDX占い】 本日の1位星座：おうし座♉ ラッキーAI：#Phi3 AIとの対話にひらめきのヒントがあるかも！ 新しい視点を取り入れてみましょう💡 #情シス #DX #企業公式つぶやき部
Top open-weight models in 2025: → Mistral 7B / Mixtral – powerful + license-free → Phi-3 – tiny, fast, mobile-native → LLaMA 3 8B – high quality, widely supported → Gemma – safe, compact OSS ≠ toy models anymore. #Mistral #LLaMA3 #Phi3 #Gemma
💰 $100 vs. $1,000,000? Agentic AI doesn’t have to cost a fortune. Discover how developers are fine-tuning, pruning, and deploying intelligent agents without big tech budgets. 🚀 Real examples. 👉 medium.com/@rogt.x1997/10… #LeanAI #DevTools #Phi3 #RunPod #Quantization
levelup.gitconnected.com
$100 vs $1 Million: How Frugal Developers Are Quietly Winning the Agentic AI Race
Inside the Rise of Lean Intelligence: Building Autonomous AI Systems That Outthink Costs 🤖 🚀
SLMs like Microsoft's Phi-3 are revolutionizing AI. Running locally on NPUs—no cloud needed—they’re fast, private, and efficient. With just 3.8B parameters, Phi-3 excels at language, code, and math. Small model, huge impact. #SLM #Phi3 #EdgeAI #OnDeviceLLM
Big brains in a small package—Microsoft’s Phi-3 AI model rivals giants, yet runs on phones and laptops. The future is efficient. ⚡️ techcrunch.com/2025/04/30/mic… #AI #Microsoft #Phi3 #MachineLearning #Innovation #Automate #Tech #Australia with SHAVIK Visit SHAVIK.ai
#Microsoft unveils Phi-3! Phi-3-mini: 3.8B parameter model trained on 3.3T tokens competitive with Mixtral 8x7B & GPT 3.5(achieves 69% on MMLU and 8.38 on MT-bench). Phi-3-small (7B,4.8T tokens) and #Phi3-medium (14B,4.8T tokens) achieves 75.3% and 78.2% on MMLU respectively.
Here's a nice GUI to test the new models #Phi3 #Llama3 #Mixtral #OpenSource #LLM yourself: @LMStudioAI. Here's how 🧵
#PHI3 'Microsoft launches PHI-3: What are SLMs, and what can they do?' :Explained by Mr Bijin Jose @bijin_jose - SLM Small Language Model #Microsoft #SLMs #LLMs #LLM #ArtificialIntelligence #GenAI #Parameters #token #Specialisation #Latency #technology #UPSC Source: IE
🥳Multi-modal Phi-3-mini is here! #LLaVA-Phi-3-mini outperforms LLaVA-v1.5-7B and matches the performance of LLaVA-Llama-3-8B in multiple benchmarks. 😊For easy applications, #GGUF weights are provided. 👉github.com/InternLM/xtuner @_akhaliq #Phi3
Exciting Update! Microsoft introduces powerful new models to the Phi-3 family on Azure. Discover more: tinyurl.com/h7nuhcpy #Phi3 #Azure #Phi3
Microsoft Phi-3 test 2: data extraction from unstructured data This #phi3 experiment is more complex: extract invoice data from a PDF invoice written in Korean. PROS: - the JSON output is perfect - it works many times (screenshot 1) - it even works with Korean language CONS: -…
Microsoft’s New Phi-3.5 Models: A Leap Forward in AI! Read Full Article : shorturl.at/aZOuN #StartupStories #Phi3 #Microsoft #AI #ArtificialIntelligence
One way to deal with #AI making up stuff is ask for a consensus :) Here's #Phi3 14b getting it wrong
Microsoft Phi-3 test 1: roleplay I am testing #phi3 , the latest small language model from @Microsoft. Test 1: roleplay. It works quite well. Ollama modelfile follows 👇 #phi3 #SLM #LLM #ollama
What do stable diffusion and Phi-3 models have in common? They can both tell you the square root of 100! 🤯 #StableDiffusion #phi3 #ai
Something went wrong.
Something went wrong.
United States Trends
- 1. GTA 6 67.2K posts
- 2. GTA VI 23.6K posts
- 3. Rockstar 56K posts
- 4. Antonio Brown 7,111 posts
- 5. Nancy Pelosi 136K posts
- 6. GTA 5 9,399 posts
- 7. Ozempic 21.4K posts
- 8. #shootingstar N/A
- 9. #bandaids 3,924 posts
- 10. #TNFonPrime 1,612 posts
- 11. Free AB N/A
- 12. Silver Slugger 5,835 posts
- 13. Paul DePodesta 2,355 posts
- 14. #LOUDERTHANEVER 1,535 posts
- 15. Fickell 1,243 posts
- 16. Rockies 4,439 posts
- 17. Grisham 1,955 posts
- 18. Justin Dean 2,169 posts
- 19. Thursday Night Football 3,793 posts
- 20. GTA 7 1,541 posts