pytorch
Official@pytorch · where the eigens are valued
Provides infrastructure orchestration for Kubernetes clusters and specialized model deployment pipelines for edge hardware targets.
Agent Skills by pytorch
Showing 34 vetted skills indexed across 3 GitHub repositories.
pytorchbinary-size
Analyze and reduce ExecuTorch runtime binary size using bloaty and size test builds.
pytorchexecutorch-kb
Searches a curated ExecuTorch knowledge base covering backends, quantization, export pitfalls, and runtime errors.
pytorchzephyr
Build and configure ExecuTorch as a Zephyr RTOS module for embedded boards.
pytorchbuilding
Builds ExecuTorch Python packages, C++ runtimes, model runners, and cross-compiled mobile libraries from source.
pytorchaoti-debug
Diagnose AOTInductor runtime errors, segfaults, and Triton index out-of-bounds failures.
pytorchmetal-kernel
Write Metal/MPS kernels for PyTorch operators on Apple Silicon.
pytorchpyrefly-type-coverage
Migrate Python files to strict Pyrefly type checking with full annotations.
pytorchscrub-issue
Fetch, reproduce, and minimize GitHub issue reproductions for PyTorch bug triage.
pytorchpt2-bug-basher
Debug PyTorch 2 compiler stack failures across Dynamo, Inductor, and AOTAutograd.
pytorchcuda-index-width
Diagnose and fix 32-bit vs 64-bit index math overflows in PyTorch CUDA kernels.
pytorchci-metrics
Query PyTorch CI workflows, test runs, and runner infrastructure metrics via Grafana.
pytorchat-dispatch-v2
Convert legacy AT_DISPATCH macros to AT_DISPATCH_V2 format in PyTorch ATen C++ code.
pytorchadd-uint-support
Add unsigned integer type support to PyTorch operator dispatch macros.
pytorchfix-issue
Reproduce, root-cause, and fix bugs reported in PyTorch GitHub issues.
pytorchtriaging-issues
Triage GitHub issues by routing, labeling, and closing usage questions.
pytorchpr-review
Review PyTorch pull requests for code quality, testing, security, and backward compatibility.
pytorchdistributed-triage
Sub-triage PyTorch distributed training issues by routing to sub-oncalls and classifying modules.
pytorchdocument-public-apis
Document undocumented PyTorch public APIs by adding Sphinx autodoc directives to doc source files.
pytorchosdc-runners-nodepools
Validate OSDC GitHub Actions runner and EKS NodePool changes with lint and test checks.
pytorchpytorch-runners-routing
Resolve PyTorch CI runner labels to final infrastructure and fleet.
pytorchosdc-harbor
Configure and troubleshoot Harbor pull-through caching for IPv6-only EKS Kubernetes clusters.
pytorchosdc-project-structure
Guides engineers in modifying the OSDC infrastructure repository while preserving its architecture.
pytorchosdc-nodelocaldns
Deploy and troubleshoot NodeLocal DNSCache in IPv6-only Kubernetes clusters.
pytorchosdc-observability
Operates and troubleshoots OSDC monitoring and logging pipelines in Kubernetes and Terraform environments.
Frequently Asked Questions About pytorch
FAQPage SchemaWhat specific tasks are enabled by these infrastructure and deployment capabilities?▼
These capabilities enable the deployment of quantized models to edge hardware, the management of IPv6-only Kubernetes clusters, and the maintenance of CI runner infrastructure. Engineers can perform model profiling, troubleshoot container registry caching, and manage cluster-level networking configurations using established infrastructure-as-code patterns.
Which target personas benefit from these infrastructure and deployment skills?▼
These skills are designed for MLOps engineers, infrastructure reliability engineers, and embedded systems developers. The content specifically supports those managing large-scale CI/CD runner fleets, Kubernetes cluster networking, and developers tasked with optimizing model execution for resource-constrained edge hardware targets.
What are the prerequisites and dependencies for implementing these deployment patterns?▼
Implementation requires an existing AWS EKS environment, Terraform for infrastructure provisioning, and Helm for package management. For model deployment, users must have the ExecuTorch runtime installed and configured with the appropriate target-specific compiler passes for hardware acceleration.