Rust + CUDA engine that streams giant MoE models from SSD, running models far larger than VRAM. Mixed-precision GGUFs. github.com/giannisanni/pulsar
Gianni Sanrochman
giannisan
AI & ML interests
Transfromers, LLMs, multimodel, vision, audio
Organizations
-
giannisan/penny-llama3-2x8b
Text Generation • 14B • Updated • 26 • 1 -
giannisan/penny5-dolphin-einstein-llama3-dare-ties-chatml
Text Generation • 8B • Updated • 20 • 1 -
giannisan/penny5-dolphin-einstein-llama3-dare-ties-chatml.gguf
8B • Updated • 35 • 1 -
mradermacher/penny-llama3-2x8b-v2-GGUF
14B • Updated • 861 • 1
Rust + CUDA engine that streams giant MoE models from SSD, running models far larger than VRAM. Mixed-precision GGUFs. github.com/giannisanni/pulsar
-
giannisan/penny-llama3-2x8b
Text Generation • 14B • Updated • 26 • 1 -
giannisan/penny5-dolphin-einstein-llama3-dare-ties-chatml
Text Generation • 8B • Updated • 20 • 1 -
giannisan/penny5-dolphin-einstein-llama3-dare-ties-chatml.gguf
8B • Updated • 35 • 1 -
mradermacher/penny-llama3-2x8b-v2-GGUF
14B • Updated • 861 • 1