Sapphire AMD Radeon RX 7900 XT 20GB GDDR6

Part of Local AI hardware: cards, computers and memory for running AI models at home.

Estimated generation speed on llama.cpp's standard benchmark (Llama 2 7B, 4-bit): 150 tokens/s.

The offload estimate assumes a PC with 64 GB of system RAM.

As described, thank you!