PromptInjection_

• Ryzen AI Max+ 495 (Gorgon Halo) with 192GB VRAM!

  Posted by PromptInjection_@reddit | LocalLLaMA | View on Reddit | 119 comments

• LLC: lightweight OpenWebUI alt - now with chat converter + custom tool calls

  Posted by PromptInjection_@reddit | LocalLLaMA | View on Reddit | 3 comments

• Qwen 3.6-27B Dense with MTP on Strix Halo Windows - Benchmarks

  Posted by PromptInjection_@reddit | LocalLLaMA | View on Reddit | 19 comments

• The "the future is fictional" problem of many local LLMs

  Posted by PromptInjection_@reddit | LocalLLaMA | View on Reddit | 54 comments

• 500k+ tokens on a 2010 laptop - how I built a chat UI that doesn't choke on large contexts

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

• LocalLightChat - the new portable lightweight Chat UI for LLMs

  Posted by PromptInjection_@reddit | LocalLLaMA | View on Reddit | 0 comments

• Gemma 4 E4B is great for short transcriptions

  Posted by PromptInjection_@reddit | LocalLLaMA | View on Reddit | 8 comments

• How to Fine-Tune LLMs on AMD Strix Halo and Other Exotic AMD Hardware

  Posted by PromptInjection_@reddit | LocalLLaMA | View on Reddit | 4 comments

• The Ultimate LLM Fine-Tuning Guide

  Posted by PromptInjection_@reddit | LocalLLaMA | View on Reddit | 8 comments

• IQ1_S Repair? What kind of crazy local AI coding hobby is this?

  Posted by PromptInjection_@reddit | LocalLLaMA | View on Reddit | 2 comments

• How "Real" Are AI Girlfriends? We Created A Unique One with a local LLM

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

• How "Real" Are AI Girlfriends? We Created A Unique with local AI

  Posted by PromptInjection_@reddit | LocalLLaMA | View on Reddit | 0 comments

• Which kind of base/fine-tunes have you done? And which data did you use?

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

• Did you know that you can use Qwen3.5-35B-A3B-Base as an instruction/reasoning Model?

  Posted by PromptInjection_@reddit | LocalLLaMA | View on Reddit | 5 comments

• Are i-Quants overrated?

  Posted by PromptInjection_@reddit | LocalLLaMA | View on Reddit | 28 comments

• Finetuned a 270M model on CPU only - full weights, no LoRA, no GPU

  Posted by PromptInjection_@reddit | LocalLLaMA | View on Reddit | 2 comments

• Are you using a SysPrompt for casual AI usage?

  Posted by PromptInjection_@reddit | LocalLLaMA | View on Reddit | 2 comments

• StatelessChatUI – A single HTML file for direct API access to LLMs

  Posted by PromptInjection_@reddit | LocalLLaMA | View on Reddit | 2 comments