Popular repositories Loading

1. spatter spatter Public

   Forked from hpcgarage/spatter

   Benchmark for measuring the performance of sparse and irregular memory access.

   C++

2. flashattn flashattn Public

   FlashAttention forward pass from scratch: PyTorch, Triton and CUDA behind one SDPA-style interface, with a KV-cache decode kernel, tests and H100 benchmarks

   Python

3. within_core within_core Public

   On-device LLM engine behind within: adapting a large language model to heterogeneous mobile hardware

   Python