smirkishere

• Locally hosted Loveable with full stack support and llama.cpp, and more

  Posted by smirkishere@reddit | LocalLLaMA | View on Reddit | 22 comments

• 4B model that looks like GPT-5 and focuses on accessibility, a11y, axe, and lighthouse

  Posted by smirkishere@reddit | LocalLLaMA | View on Reddit | 12 comments

• Open Source Lovable with Custom Agents, Full Stack Support, and Local Models

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

• 4B model RL trained on a11y, lighthouse, and axe scores to generate more accessible websites

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

• WEBGEN, UIGEN-FX, UIGENT research preview releases

  Posted by smirkishere@reddit | LocalLLaMA | View on Reddit | 12 comments

• 4B models are consistently overlooked. Runs Locally and Crushes It. Reasoning for UI, Mobile, Software and Frontend design.

  Posted by smirkishere@reddit | LocalLLaMA | View on Reddit | 80 comments

• WEBGEN-OSS Web Design Model - a model that runs on a laptop and generates clean responsive websites from a single prompt

  Posted by smirkishere@reddit | LocalLLaMA | View on Reddit | 44 comments

• WEBGEN-4B: Quality Web Design Generation

  Posted by smirkishere@reddit | LocalLLaMA | View on Reddit | 39 comments

• Inference optimizations on ROCM?

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

• We built a visual drag-n-drop builder for multi-agent LLM Orchestration (TFrameX + Agent Builder, fully local, MIT licensed)

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

• UIGEN-X-0727 Runs Locally and Crushes It. Reasoning for UI, Mobile, Software and Frontend design.

  Posted by smirkishere@reddit | LocalLLaMA | View on Reddit | 77 comments

• Is it possible to run 32B model on 100 requests at a time at 200 Tok/s per second?

  Posted by smirkishere@reddit | LocalLLaMA | View on Reddit | 33 comments