flashrt/fp8-gemm
Updated
AI & ML interests
training & inference infra
Recent Activity
flashrt/fp4-gemm
Updated
flashrt/masked-mha-runtime
Updated
flashrt/fp4-fused-ops
Updated • 321
flashrt/transformer-fused-ops
Updated
flashrt/flashrt-qkv-cache-rope
Updated
flashrt/flashrt-adaptive-norms
Updated
flashrt/flashrt-fp8-ffn
Updated
flashrt/flashrt-residual-norm-quant
Updated
flashrt/gated-delta-attention
Updated • 234
flashrt/grouped-moe-gemv
Updated • 101
flashrt/transformer-layout-primitives
Updated
flashrt/diffusion-step-ops
Updated
flashrt/fa4-cute-runtime
Updated
flashrt/world-model-conv
Updated
flashrt/weight-only-ffn
Updated
flashrt/vl-transformer-primitives
Updated
flashrt/fp8-cross-attention-blackwell
Updated
flashrt/flashrt-vla-residual-gates
Updated
flashrt/flashrt-spatiotemporal-layout
Updated
flashrt/flashrt-gemm-epilogues
Updated
flashrt/flashrt-fused-quant
Updated
flashrt/flashrt-fp8-swiglu-ffn
Updated
flashrt/blockwise-fp8-producers
Updated
flashrt/bf16-linear-gemv
Updated
flashrt/adaptive-layernorm-producers
Updated
flashrt/grouped-moe-gemm
Updated
flashrt/fp8-kv-attention
Updated
flashrt/smallm-ffn-megakernels-blackwell
Updated
flashrt/flashrt-nvfp4
Updated