NickNau

• 'Token Counter' app to get number of tokens in local files. Supports various models.

  Posted by NickNau@reddit | LocalLLaMA | View on Reddit | 14 comments

• [TEST] Prompt Processing VS Inferense Speed VS GPU layers

  Posted by NickNau@reddit | LocalLLaMA | View on Reddit | 20 comments

• LLM as survival knowledge base

  Posted by NickNau@reddit | LocalLLaMA | View on Reddit | 152 comments

• Speculative decoding for on-CPU MoE?

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

• Speculative decoding can identify broken quants?

  Posted by NickNau@reddit | LocalLLaMA | View on Reddit | 124 comments

• Proxy to get Claude in Open WebUI

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

• Qwen2.5 1M context works on llama.cpp?

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

• [FINAL TEST] Power limit VS Core clock limit efficiency

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

• Will we continue to tolerate politic bots?

  Posted by NickNau@reddit | LocalLLaMA | View on Reddit | 192 comments

• GPU Speed vs Tokens per Second & Power Draw [test results]

  Posted by NickNau@reddit | LocalLLaMA | View on Reddit | 35 comments

• How to run DeepSeek V2.5 quants?

  Posted by NickNau@reddit | LocalLLaMA | View on Reddit | 15 comments

• Need advice on 6x3090 inference software setup

  Posted by NickNau@reddit | LocalLLaMA | View on Reddit | 27 comments

• Need advice on 4x3090 inference build

  Posted by NickNau@reddit | LocalLLaMA | View on Reddit | 52 comments

• True random answer from LLM

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

• bartowski vs TheDrummer etc

  Posted by NickNau@reddit | LocalLLaMA | View on Reddit | 23 comments

• DDR5 2x32 + 2x48 - will it work?

  Posted by NickNau@reddit | buildapc | View on Reddit | 9 comments