7 projects
cubemind
Neuro-vector-symbolic architecture for compositional reasoning on consumer hardware
grilly
GPU-accelerated neural network operations using Vulkan compute shaders
optimum-grilly
HuggingFace Optimum backend for Grilly — Vulkan GPU inference on any GPU
grillycompression
Activation, KV-cache, and communication compression pipelines — optional grilly extension
grillydistil
Temperature-scaled distillation with SA-KD — optional grilly extension
grillyoptimum
HuggingFace Optimum-compatible Vulkan backend — optional grilly extension
grillyinference
Native fp16 inference engine for Llama models — optional grilly extension