2 مستودعات
Reconstructing a compiled kernel adapter from previously cached compilation results to avoid recompilation.
Distinct from Kernel Caching Systems: More specific than Kernel Caching Systems: focuses on loading from cache, not general persistence.
Explore 2 awesome GitHub repositories matching artificial intelligence & ml · Cached Kernel Loading. Refine with filters or upvote what's useful.
Voilà is a tool that converts Jupyter notebooks into standalone interactive web applications. It renders notebook cells as HTML web components, preserving live widgets while stripping source code by default, and gives each viewer a dedicated Jupyter kernel for isolated widget state and callback execution. The project runs as a Jupyter server extension, reusing existing server infrastructure for notebook serving and authentication. It supports directory-based notebook hosting, serving all notebooks in a folder as a browsable collection of web applications from a single command. Voilà also prov
Starts a notebook's kernel before the first user request so the dashboard loads faster.
TileLang is a Python-embedded domain-specific language compiler that JIT-compiles and autotunes GPU kernels. It uses a tile-based DSL, automatic software pipelining, and parallel autotuning to generate optimized GPU kernels at runtime. It supports tensor core operations with Pythonic syntax, automatic memory management, and thread mapping. The compiler searches over tile sizes, thread counts, and scheduling policies, compiling and benchmarking candidates in parallel to find the fastest kernel. It also caches compiled binaries and tuning results to disk for reuse across sessions. TileLang inc
Loads compiled kernels from cache to avoid recompilation across sessions.