flashrt/transformer-fused-ops
Updated
AI & ML interests
training & inference infra
Recent Activity
View all activity
flashrt/flashrt-qkv-cache-rope
Updated
flashrt/flashrt-spatiotemporal-layout
Updated
flashrt/fp8-gemm
Updated
flashrt/adaptive-layernorm-producers
Updated
flashrt/padded-fp8-producers
Updated
flashrt/flashrt-vla-residual-gates
Updated
flashrt/fp8-kv-attention
Updated
flashrt/fp4-fused-ops
Updated
flashrt/blockwise-fp8-producers
Updated
flashrt/diffusion-step-ops
Updated
flashrt/world-model-conv
Updated
flashrt/transformer-layout-primitives
Updated
flashrt/fp8-prefill-attention-blackwell
Updated • 1
flashrt/grouped-moe-gemm
Updated
flashrt/smallm-ffn-megakernels-blackwell
Updated
flashrt/small-matrix-cholesky
Updated
flashrt/fa2-seqused-runtime
Updated
flashrt/flashrt-fp8-swiglu-ffn
Updated
flashrt/flashrt-fp8-ffn
Updated
flashrt/fp4-gemm
Updated
flashrt/weight-only-ffn
Updated
flashrt/int4-blackwell
Updated • 1
flashrt/flashrt-flex-attention-train
Updated
flashrt/vl-transformer-primitives
Updated
flashrt/turboquant-kv
Updated
flashrt/speculative-draft-primitives
Updated
flashrt/sageattention2-blackwell
Updated
flashrt/linear-attention-seq-state
Updated
flashrt/linear-attention-primitives
Updated