AMD
Official@amd
Offers specialized quantization and model optimization capabilities for deploying high-performance transformer architectures on ROCm-enabled hardware and ONNX runtimes.
Agent Skills by AMD
Showing 31 vetted skills indexed across 1 GitHub repositories.
amdquark-onnx-autosearch-pro
Plans and orchestrates ONNX quantization auto-search runs with Optuna-driven presets and trial budgets.
amdquark-torch-llm-ptq-eval
Orchestrate post-training quantization workflows for Torch LLMs with validation and accuracy evaluation.
amdquark-onnx-doc-drift-check
Compare Quark ONNX skill guidance against upstream docs and source.
amdquark-onnx-skill-sync
Detect upstream AMD Quark ONNX drift in skill guidance and contracts.
amdquark-onnx-eval-runner
Verify Quark ONNX skill routing, planning, artifacts, and recovery checks.
amdquark-torch-eval-runner
Verify Quark skill routing, planning, artifact, and recovery outputs.
amdquark-torch-skill-sync
Detects upstream Quark changes that invalidate skill instructions or contracts.
amdquark-torch-doc-drift-check
Compare Quark skill guidance against upstream docs and source to detect outdated commands, flags, and options.
amdquark-skill-creator
Create and restructure Quark Agent Skills to meet format and governance contracts.
amdquark-workspace-validate
Validate workspace paths, model references, and output directories before Quark workflows.
amdquark-env-preflight
Collect OS, Python, GPU, and accelerator facts for Quark installation readiness.
amdquark-onnx-ptq-workflow
Guide end-to-end post-training quantization of ONNX models with AMD Quark.
amdquark-torch-llm-ptq-workflow
Orchestrate AMD Quark post-training quantization workflows for PyTorch and ONNX models.
amdquark-onnx-quant-plan
Build hardware-aware Quark ONNX quantization plans from model analysis and deployment intent.
amdquark-onnx-result-validator
Validate Quark ONNX quantization outputs by checking auxiliary files, initializers, metadata, and QDQ signals.
amdquark-onnx-install
Install and verify ONNX Runtime stacks for AMD Quark workflows.
amdquark-onnx-model-intake
Extracts ONNX model metadata needed for Quark PTQ planning and compatibility assessment.
amdquark-onnx-router
Routes Quark ONNX user requests to the smallest correct downstream workflow.
amdquark-onnx-debug
Diagnose Quark ONNX installation, calibration, quantization, and custom-op failures.
amdquark-torch-llm-eval
Evaluate LLM accuracy on AMD ROCm with vLLM, SGLang, or ATOM serving.
amdquark-torch-result-validator
Validate Quark quantization outputs via file, tensor, and config checks.
amdquark-torch-install
Install and verify PyTorch builds for CUDA, ROCm, or CPU backends.
amdquark-torch-router
Route AMD Quark user intents to the correct downstream skill or workflow.
amdquark-torch-model-intake
Analyze HuggingFace model architectures for Quark PTQ planning readiness.
Frequently Asked Questions About AMD
FAQPage SchemaWhat specific model optimization tasks are supported?▼
These capabilities enable end-to-end post-training quantization for PyTorch and ONNX models. Users can perform model intake analysis, generate hardware-aware quantization plans, execute file-to-file checkpoint compression, and validate final artifacts against metadata and tensor-level accuracy requirements.
Which technical personas benefit from these capabilities?▼
Machine learning engineers and deployment specialists focused on optimizing transformer models for AMD ROCm hardware benefit most. These capabilities assist those managing model drift, verifying environment readiness, and ensuring quantization outputs meet strict deployment contracts.
What are the primary prerequisites for implementation?▼
Implementation requires a configured environment with PyTorch or ONNX Runtime stacks, alongside ROCm drivers for hardware acceleration. Users must validate workspace paths, model references, and system-level dependencies using preflight checks before initiating quantization or evaluation sequences.