ifioravanti

• Apple M3 Ultra 512GB vs NVIDIA RTX 3090 LLM Benchmark

  Posted by ifioravanti@reddit | LocalLLaMA | View on Reddit | 57 comments

• Apple MLX Quantizations Royal Rumble 🔥

  Posted by ifioravanti@reddit | LocalLLaMA | View on Reddit | 9 comments

• Ollama performance on M2 Ultra - M3 Max - Windows Nvidia 3090 and WSL2 Nvidia 3090

  Posted by ifioravanti@reddit | LocalLLaMA | View on Reddit | 7 comments

• 🔥 DeepSeek R1 671B Q4 - M3 Ultra 512GB with MLX🔥

  Posted by ifioravanti@reddit | LocalLLaMA | View on Reddit | 209 comments

• This M2 Ultra v2 M3 Ultra benchmark by Matt Tech Talks is just wrong!

  Posted by ifioravanti@reddit | LocalLLaMA | View on Reddit | 67 comments

• Llama 405B running locally!

  Posted by ifioravanti@reddit | LocalLLaMA | View on Reddit | 63 comments

• DeepSeek R1 Distill Qwen 7B Q4 large context (up to 128K) tests

  Posted by ifioravanti@reddit | LocalLLaMA | View on Reddit | 13 comments

• LM Studio - Hugging Face Model Manager

  Posted by ifioravanti@reddit | LocalLLaMA | View on Reddit | 6 comments

• 🤔 Can the iogpu.wired_limit_mb setting on MacOS be persisted? 📣 YES!

  Posted by ifioravanti@reddit | LocalLLaMA | View on Reddit | 6 comments

• Best LLM with large context window? Not for coding in Feb 2024

  Posted by ifioravanti@reddit | LocalLLaMA | View on Reddit | 1 comments

• MLX now supports loading GGUF directly from Huggingface models

  Posted by ifioravanti@reddit | LocalLLaMA | View on Reddit | 1 comments