skills.rest
ulf1
ulf1COMMUNITY
Explore all skills by this author →
GitHub

323-pytorch-unittest

Provides linting and auto-fix tooling for Python 3.6+ to ensure compatibility.

Updated May 21, 2026
Search Tags:#pytest#unit-testing#pytorch#determinism#autograd#torch.testing.assert_close#device-placement
One-click install
npx skills add https://github.com/ulf1/trading-regime --skill 323-pytorch-unittest
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: 323-pytorch-unittest
Source: https://github.com/ulf1/trading-regime/tree/main/.agent/skills/323-pytorch-unittest
Command: npx skills add https://github.com/ulf1/trading-regime --skill 323-pytorch-unittest

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

This Skill prevents flaky or misleading unit tests for PyTorch by enforcing correct tensor comparisons, tolerance handling, dtype/shape/device checks, deterministic RNG usage, and gradient-flow validation.

Core Features & Use Cases

  • Robust Tensor Assertions: Use torch.testing.assert_close with explicit rtol and atol instead of unsafe == checks for tensors, avoiding incorrect boolean semantics and brittle equality failures.
  • Invariant Validation: Assert shapes, dtypes, and device placement so silent upcasts, reshaping errors, or CPU/GPU mismatches are caught immediately.
  • Autograd & Determinism Checks: Verify gradients after backward() and ensure deterministic behavior via controlled seeding for reproducible results.
  • Use Cases: Testing tensor math correctness, validating custom layers and loss functions, checking training-loop gradients, and ensuring model outputs are stable across refactors.

Quick Start

Use the 323-pytorch-unittest skill when writing or reviewing any pytest/unittest that imports torch so your assertions verify values, shapes/dtypes, device placement, gradients, and determinism with appropriate tolerances.

Frequently Asked Questions about 323-pytorch-unittest

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I assert tensor equality in pytest without causing flaky tests?▼

Use torch.testing.assert_close with explicit rtol and atol tolerances instead of unsafe == checks to prevent brittle equality failures and incorrect boolean semantics when comparing PyTorch tensors in pytest.

How do I validate autograd gradients flow correctly in PyTorch unit tests?▼

Validate autograd gradients in PyTorch unit tests by calling backward and verifying the resulting gradients, ensuring custom layers and loss functions maintain proper gradient flow throughout the training loop.

What is the best way to check tensor shape, dtype, and device placement during testing?▼

Checking tensor shape, dtype, and device placement requires explicit invariant assertions to catch silent upcasts, reshaping errors, and CPU or GPU mismatches immediately during PyTorch unit testing.

How do I make PyTorch test behavior deterministic across CPU and GPU contexts?▼

Make PyTorch test behavior deterministic across CPU and GPU contexts by implementing controlled RNG seeding, ensuring reproducible results and eliminating random number generation variance in your test suite.

Why does using == for tensor comparisons fail in PyTorch testing?▼

Using == for tensor comparisons fails in PyTorch testing because it triggers incorrect boolean semantics and brittle equality failures, requiring torch.testing.assert_close with documented tolerances instead.

Can I use this approach for testing custom PyTorch layers and loss functions?▼

Yes, this testing approach applies to validating custom layers and loss functions by enforcing value correctness, shape and dtype invariants, device placement, and verified gradients within pytest or unittest scenarios.

Related Skills

Explore More Software Engineering Skills→
cheetah-alocheetah-alo

pytest

Run unittest-based Python test suites with pytest for filtering and coverage.

Community
Basic
plurigridplurigrid
60

qa-regression

Automate Playwright-based QA regression tests for login, dashboard, and user management flows.

Official
Advanced
dralgorhythmdralgorhythm
111

optimizing-code

Profile hot code paths, implement caching, and benchmark before and after.

Community
Advanced
GhostOf0daysGhostOf0days

financial-computing-numpy

Implement vectorized NumPy pipelines for financial computations with diagnostics.

Community
Advanced