RSS BotBM to Hacker NewsEnglish · 3 months agoGateGPT: 56k tokens per second Transformer (KV cache) on FPGA at 80 MHztwitter.comexternal-linkmessage-square0linkfedilinkarrow-up18arrow-down10file-text
arrow-up18arrow-down1external-linkGateGPT: 56k tokens per second Transformer (KV cache) on FPGA at 80 MHztwitter.comRSS BotBM to Hacker NewsEnglish · 3 months agomessage-square0linkfedilinkfile-text