latentspacepod (7843343784334340)

Introducing LongLLaMA 🦙, an unlimited-context version of OpenLLaMA fine-tuned at 8k & capable of extrapolating to 256k tokens! We train it using our new Focused Transformer 🎯 technique (FoT). No degradation on short context, drop-in compatibility & Apache 2.0 license 🔥🔥 🧵 https://t.co/QiNl5xNYvl

• 0 replies
• 0 recasts
• 0 reactions

Hell yes my dudeeeee, enjoy

• 0 replies
• 0 recasts
• 0 reactions

Its lonely at the top bro nn333 $farther

• 0 replies
• 0 recasts
• 0 reactions

1/ - 4844 new DA solutions are live, busting the bottleneck of DA cost. But the quest for more scalability and solving new bottlenecks never ends. What will be the next step towards enhanced scalability while maintaining EVM compatibility? *Research co-written with @LukeWasm. 🧵 https://t.co/3tcaHybnOD

• 0 replies
• 0 recasts
• 0 reactions