skill
职业分类
描述
更新
inferencex-optimize
数据科学家
Run the InferenceX benchmark/profiling workflow for a config key. When the user names a model or config key, immediately start a batched guided setup, then execute.
2026-04-01
vllm-optimize
数据科学家软件开发工程师
Run vLLM benchmark and profiling workflow in containerized environments. Provides automated model loading, concurrency sweep, torch profiling, and GPU kernel analysis. Works with AMD MI355X/MI300X and NVIDIA GPUs.
2026-03-31