AMD-AGI/Hyperloom

inference-optimizer

Launches and monitors Hyperloom's multi-agent inference optimizer for LLM serving on AMD GPUs. Use when the user asks to optimize an inference model, run Magpie benchmarks/profiles, resume an inference_optimizer session, tune SGLang/vLLM serving parameters,…

原文语言：英语

session-breakdown

Build a single self-contained `session_breakdown.json` capturing every fact a dashboard needs about one hyperloom optimization session. Use when the user mentions session-breakdown, kernel attribution, a stats/reporting service, capability summary, or wants…

原文语言：英语

robustness-agent

Independent guardian daemon for Hyperloom inference optimization. Implements the inference_optimizer "robustness" reactor so the Coordinator can call it as a Backend, plus a standalone loop for dev. Owns continuous health monitoring, RCA, and…

原文语言：英语

hyperloom-custom-advanced

Run an advanced configurable Hyperloom optimization session with explicit model, framework, workload, objective, and phase toggles. Use when the user wants more control than the fixed 3h or 12h demo presets.

原文语言：英语

hyperloom-qwen3-8b-3h

Run a 3-hour Hyperloom Qwen3-8B optimization session without the Kernel Agent. Use when the user wants a short, no-kernel Hyperloom demo on the local AMD ROCm environment.

原文语言：英语

critic-agent

Critic layer for the inference optimizer. Use when Conductor asks for a Critic Review verdict on Orchestration or Kernel proposals, conversation-driven decision review, KB recall/ingest guidance, cross-run synthesis, or Devil's advocate review signals.

原文语言：英语

hyperloom-qwen3-14b-fp8-12h

Run a 12-hour Hyperloom Qwen3-14B-FP8 optimization session. Use when the user wants a medium-length Hyperloom demo on the local AMD ROCm environment.

原文语言：英语

hyperloom-remote-mn-qwen3-30b

Run a 4-hour multi-node Hyperloom Qwen3-30B-A3B optimization (Infera PD-disaggregated or RayJob aggregated) with --nodes 2 and sglang MoE tuning on MI325X. Hand this skill to the agent to launch and monitor the run.

原文语言：英语

hyperloom-inference-optimizer-multi-node

Multi-node companion to the inference_optimizer skill. Use when the user prompt asks for inference optimization that needs more GPU / memory than a single pod provides (i.e. ``nodes >= 2``) — typical prompt signals are ``Nodes=N`` / ``N pods`` / ``TP=N``…

原文语言：英语

kernel-trace-analysis

Profile GPU kernels using rocprofv3 to collect ATT instruction-level traces, then analyze the trace data using hotspot_analyzer.py to identify top-K stall hotspots (VMEM-load, VMEM-wait, LDS/SMEM-wait, barrier, MFMA stalls) mapped back to source lines, and…

原文语言：英语

quantization-agent

Hyperloom sub-agent that drives the AMD Quark PTQ skill chain end-to-end from a natural-language prompt. Produces a HuggingFace-format quantized model directory plus a structured assessment of every artifact that the Quark workflow + validator + llm-eval…

原文语言：英语

hyperloom-setup

Configure Hyperloom in the current agent workspace after pip install --target . by collecting LLM settings, choosing a bare-metal or Docker run mode, writing .env, and running the setup backend on baremetal hosts only.

原文语言：英语