Requisitos de hardware
#1
by akudamcs - opened
Qué hardware necesito para hacer uso del modelo en local?
Hi! To run AIkuda locally you'll need:
Minimum: 1× GPU with 24 GB VRAM (RTX 3090, RTX 4090, RTX 5090) running Q4 quantization (17 GB) via vLLM or Ollama.22 GB) for better quality.
Recommended: 1× GPU 24 GB with Q6 quantization (
Optimal: 3× GPUs with 16 GB each (48 GB total) running full BF16 with tensor parallelism.
CPU only: Possible with llama.cpp + GGUF Q4, but requires 64 GB RAM and will be slow.
We'll publish GGUF quantized versions alongside the main model for easier local deployment.
Stay tuned! 🛡️