annealing-inversion
Community@annealing-inversion
Agent Skills by annealing-inversion
Showing 23 vetted skills indexed across 1 GitHub repositories.
annealing-inversionsglang-diffusion-modelopt-quant
Quantize diffusion DiT models with NVIDIA ModelOpt for SGLang Diffusion.
annealing-inversionsglang-diffusion-benchmark-profile
Benchmark diffusion denoise latency and profile SGLang native backend bottlenecks.
annealing-inversionsglang-diffusion-add-model
Integrate new diffusion models into the SLang framework via standard stages.
annealing-inversionsglang-diffusion-ako4all-kernel
Automate AKO4ALL-based optimization of SGLang diffusion kernels with benchmarking and validation.
annealing-inversionsglang-diffusion-performance
Benchmark diffusion configurations across SGLang and Diffusers backends.
annealing-inversionclean-startup-log
Suppress third-party warnings and deprecation notices in SGLang server startup logs.
annealing-inversionwrite-sglang-test
Create and register SGLang CI/UT tests using CustomTestCase templates.
annealing-inversionsglang-cherrypick
Automate batch cherry-picking of merged PRs onto a release branch via GitHub Actions.
annealing-inversionci-workflow-guide
Design GitHub CI workflows with stage ordering, gating, and fast-fail.
annealing-inversionlarge-class-init-style
Enforce init_* helper patterns in Scheduler, TokenizerManager, and ModelRunner constructors.
annealing-inversionspeculative-naming
Standardize speculative decoding identifier names in the SGLang codebase.
annealing-inversionadd-jit-kernel
Automate creation of JIT CUDA kernels for SGLang with bindings and tests.
annealing-inversiondebug-cuda-crash
Enable Kernel API logging to diagnose CUDA crashes in SGLang.
annealing-inversionllm-serving-auto-benchmark
Benchmark LLM serving frameworks to identify the best deployment command.
annealing-inversionsglang-prod-incident-triage
Replay live SGLang serving faults into reproducible debug paths.
annealing-inversiongenerate-profile
Profile an SGLang server run and generate a Chrome trace and server_args.json.
annealing-inversionadd-sgl-kernel
Add AOT CUDA/C++ kernels to sgl-kernel with tests and benchmarks.
annealing-inversionllm-torch-profiler-analysis
Analyze LLM torch-profiler traces into kernel, overlap, and fuse pattern reports.
annealing-inversionsglang-bisect-ci-regression
Bisect commit windows and correlate CI failures with runners for SGLang.
annealing-inversionsglang-sota-performance
Benchmark SGLang against vLLM and TensorRT-LLM, then patch bottlenecks.
annealing-inversiondebug-distributed-hang
Diagnose distributed-inference hangs in SGLang across parallel setups using per-rank traces.
annealing-inversionenv-var-conventions
Centralize SGLANG_* environment variables in Envs descriptors with typed access and deprecation.
annealing-inversionmechanical-refactor-verify
Verify mechanical refactors via transform scripts against target commits.