Threads @krishna.dxx — DcveYxpH4up
Someone got Kimi K3😭 a 2.78 TRILLION parameter model running on a CPU with 8GB of RAM. No GPU. No framework. No BLAS. Just 176KB of C code and a disk. More RAM only makes it faster. The output never changes. Your normal laptop can now run the same model as a $50K workstation. It just takes longer. Link in next reply 🔗 Follow so you don’t miss the next one 👀 GitHub - FareedKhan-dev/kimi-k3-in-c: A 2.78-trillion-parameter Kimi K3 running inference on a single CPU in 8.24 GB of RAM. Portable C99: no BLAS, no framework, no GPU.