DeepSeekAMD
DeepSeek's Top Model Runs on Two AMD GPUs — Local AI No Longer NVIDIA's Solo Act
Reddit developer neuroma ran DeepSeek-V4-Flash at 40-50 tok/s on two AMD Radeon R9700 cards with 262K context. Top local open-source AI no longer need
Sep 23·2 min read
