pytorch_template CLAUDE.md

Project instructions for a PyTorch template, a reusable starting codebase for machine-learning projects. They describe development steps and rules for keeping migration documentation current.

In plain words
What is it for?
Adding or changing features, updating version-detection tables and migration notes, and following the project's Git and development workflow.
Why use it?
They keep template changes compatible with the migration process used to move projects to newer template versions.

Instructions file

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add instructions/axect/pytorch_template/claude-md
Clone the repo
git clone --depth 1 https://github.com/Axect/pytorch_template
Per session 331 This file is loaded in full into every session.
When invoked 331 The same file — it is already loaded in full.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00331 $0.00331
Opus 5 $0.00166 $0.00166
Sonnet 5 $0.00066 $0.00066
Haiku 4.5 $0.00033 $0.00033

Measured 2d ago against content hash 3ab7bbe47737, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

pytorch_template CLAUDE.md scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

CLAUDE.md · 31 lines

What it actually says

pytorch_template

Migration Skill Maintenance Rule

When modifying any of these template source files, you MUST also update the migration skill documentation:

Tracked source files:

  • callbacks.py, config.py, util.py, cli.py
  • checkpoint.py, pruner.py, provenance.py, main.py

Migration docs to update:

  • skills/pytorch-migrate/SKILL.md (version detection table + migration summary)
  • skills/pytorch-migrate/references/migrations.md (detailed migration steps)

When adding a new feature:

  1. Create a new migration section (M7, M8, ...) in references/migrations.md
  2. Add detection marker(s) to the version table in SKILL.md
  3. Include: detection grep pattern, which files/classes/methods to copy from $TEMPLATE_DIR, wiring instructions
  4. Do NOT embed code snippets — the migration clones the latest template and reads code from there

When modifying an existing feature:

  • Update the relevant migration section's structural description (class names, method signatures, field names) if they changed
  • The actual code is always read from the template clone, so minor code changes don't require migration doc updates

Development Workflow

  • Use Gitflow: feature branches from main
  • Run pytest tests/ -x -q before committing
  • Run python -m cli preflight configs/run_template.yaml --device cpu to verify template health
  • A pre-push hook enforces migration doc updates when source files change
Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 2d ago First seen · 31 lines · 331 tokens per session scan A 3ab7bbe47737

Subscribe to this mod's changes

pytorch_template CLAUDE.md is an instructions file published in the GitHub repository Axect/pytorch_template (10 stars, last pushed 3mo ago), licensed MIT. It adds 331 tokens to every session, about $0.0017 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.

Related

Other instructions, from other repositories