What are your prompts for code assitants?
How To Run Deepseek R1 671b Fully Locally On a $2000 EPYC Server
Ollama is confusing people by pretending that the little distillation models are "R1"
Almost a year later, I can finally do this. A small teaser of a project I'm working on
48gb vs 96gb VRAM for fine-tuning
6x AMD Instinct Mi60 AI Server vs Llama 405B + vLLM + Open-WebUI - Impressive!
Asking for hardware recommendations for a personal machine capable of running +70B models. With cloud options I have to re-download the model every time. Should I bite the bullet and get Mac Studio M2 Ultra ($7000 after tax), or build a PC? What specs do you recommend?
Now that Phi-4 has been out for a while what do you think?
Phi 4 is just 14B But Better than llama 3.1 70b for several tasks.
Wow this maybe probably best open source model ?
To understand the Project DIGITS desktop (128 GB for 3k), look at the existing Grace CPU systems
For those who care about how o1 works technically, OpenAI has stated that o1 was built using reinforcement fine-tuning, which was announced by OpenAI on December 6 as day 2 of Shipmas
UwU 7B Instruct
My personal guide for developing software with AI assistance
Llama 3.3 70B Help
How long until AI agent that interact with email, calendar, to-do list, etc?
I've tried Granite 3.1 3b. It was very fast and very bad.
Doing the thing: speculations about the next DBRX release?
Budget AKA poor man Local LLM.
Intel preparing Arc (PRO) "Battlemage" GPU with 24GB memory - VideoCardz.com
Marco-o1 posted on ollama
PyTorch just released their own llm solution - torchchat
Nvidia just dropped its Multimodal model NVLM 72B
Linux Compatibility w/ Intel Ultra Processors
Would anyone be interested in buying pre-built and pre-setup 4x3090/4090 or 8x3090/4090 PCs for training/inference?
