minus-squareLongjumping-Bake-557@alien.topBtoLocalLLaMA@poweruser.forum•X.AI Grok could potentially be open sourced on a 6 month delay from launchlinkfedilinkEnglisharrow-up1·1 year agoWhich is hilarious seeing how he kept joking about “closedai” for a while linkfedilink
Longjumping-Bake-557@alien.topB to LocalLLaMA@poweruser.forumEnglish · 1 year agoWhy is no one releasing 70b models?plus-squaremessage-squaremessage-square35fedilinkarrow-up11arrow-down10
arrow-up11arrow-down1message-squareWhy is no one releasing 70b models?plus-squareLongjumping-Bake-557@alien.topB to LocalLLaMA@poweruser.forumEnglish · 1 year agomessage-square35fedilink
Longjumping-Bake-557@alien.topB to LocalLLaMA@poweruser.forumEnglish · 1 year agoDiscrepancy between TheBloke_Orca-2-13B-GPTQ and the original one with the tested logic questionalien.topimagemessage-square2fedilinkarrow-up11arrow-down10
arrow-up11arrow-down1imageDiscrepancy between TheBloke_Orca-2-13B-GPTQ and the original one with the tested logic questionalien.topLongjumping-Bake-557@alien.topB to LocalLLaMA@poweruser.forumEnglish · 1 year agomessage-square2fedilink
minus-squareLongjumping-Bake-557@alien.topBtoLocalLLaMA@poweruser.forum•NVidia H200 achieves nearly 12,000 tokens/sec on Llama2-13B with TensorRT-LLMlinkfedilinkEnglisharrow-up1·1 year agoAnd that’s on a die just slightly bigger than the 4090. Unless they increased the size compared to h100? linkfedilink
Longjumping-Bake-557@alien.topB to LocalLLaMA@poweruser.forumEnglish · 1 year agoSynthIA 70b on a single 3090, 0.17tokens/splus-squaremessage-squaremessage-square1fedilinkarrow-up11arrow-down10
arrow-up11arrow-down1message-squareSynthIA 70b on a single 3090, 0.17tokens/splus-squareLongjumping-Bake-557@alien.topB to LocalLLaMA@poweruser.forumEnglish · 1 year agomessage-square1fedilink
Which is hilarious seeing how he kept joking about “closedai” for a while