New: 98,000+ skills indexed from GitHub repos with 100+ stars · every skill safety-checkedBrowse top skillsTop skills
Esc to closeBrowse all skills
Submit a skillBrowse skills
Submit a skillBrowse skills

Agent skill

Anomalib Tiled Ensemble

by open-edge-platform in open-edge-platform/anomalib

Runs and configures the anomalib tiled-ensemble pipeline, which trains/evaluates one model per image tile and merges results (with optional seam smoothing) for high-resolution anomaly detection.

Apache-2.0Auto-check passedData & Analytics

Install Anomalib Tiled Ensemble

skills CLI
$ npx skills add open-edge-platform/anomalib --skill anomalib-tiled-ensemble -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI
$ gh skill install open-edge-platform/anomalib anomalib-tiled-ensemble --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy
$ git clone --depth 1 https://github.com/open-edge-platform/anomalib.git skills-src && mkdir -p .claude/skills && cp -r skills-src/.agents/skills/anomalib-tiled-ensemble .claude/skills/anomalib-tiled-ensemble && rm -rf skills-src

Use ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.

Claude Code skills documentation · loads skills from .claude/skills/

Facts

Skill name
anomalib-tiled-ensemble
GitHub stars
6.2k
Token cost
~1.4k tokens
SKILL.md length
435 words
Files
2
Skills in repo
17
Repo updated
First seen
Licence
Apache-2.0

At a glance

Runs and configures the anomalib tiled-ensemble pipeline, which trains/evaluates one model per image tile and merges results (with optional seam smoothing) for high-resolution anomaly detection.

  • The user wants to train with image tiling
  • SKILL.md covers Code locations, Running it, Config structure and Gotchas, plus 1 more section
  • Calls python
  • Mentions tiled ensemble

What it does

Anomalib Tiled Ensemble is an agent skill from open-edge-platform/anomalib. Runs and configures the anomalib tiled-ensemble pipeline, which trains/evaluates one model per image tile and merges results (with optional seam smoothing) for high-resolution anomaly detection. Use when the user wants to train with image tiling, mentions "tiled ensemble", or needs to tune tiling/stride/seam-smoothing config. Do not use for regular single-model training (see anomalib-training) or the multi-model benchmarking pipeline (see anomalib-benchmarking).

Its SKILL.md is about 1.4k tokens, which your agent loads only when the skill is triggered. The skill folder holds 2 other files (for example `evals/evals.json`).

It sits in Data & Analytics, covering Threat modeling, Anomaly detection and Fine-tuning. The repository describes itself as: An anomaly detection library comprising state-of-the-art algorithms and features such as experiment management, hyper-parameter optimization, and edge inference. The licence is Apache-2.0.

When your agent uses it

  • The user wants to train with image tiling
  • Mentions tiled ensemble
  • Needs to tune tiling/stride/seam-smoothing config
  • Regular single-model training (see anomalib-training)

Example prompts

  • “tiled ensemble”
  • “/anomalib-tiled-ensemble”

Requirements

  • Python 3

What it can do on your machine

Read from SKILL.md and the folder at commit dc087d5. It shows what the files ask for, not the result of running them.

  • Tool permissions

    Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

    From allowed-tools in the SKILL.md frontmatter.

  • Runs code

    Shell commands in SKILL.md call:

    • python

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    No URLs in SKILL.md.

    From URLs in SKILL.md, links to its own repository left out.

  • Credentials

    Names no API keys, tokens, secrets or passwords.

    From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Anomalib Tiled Ensemble loads about 1.4k tokens when it runs. Until then it costs about 123 tokens; SKILL.md has 435 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~123
When it runs · the whole SKILL.md, loaded when a task matches
~1.4k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check passed

The automated check found no risky patterns in SKILL.md.

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from open-edge-platform/anomalib at commit dc087d5, republished under its Apache-2.0 licence (© open-edge-platform). 435 words, ~1,351 tokens.

Download SKILL.mdSave it as .claude/skills/anomalib-tiled-ensemble/SKILL.md (or your agent's skills folder). This skill also uses 1 other file; get the full folder from GitHub.
name
anomalib-tiled-ensemble
description
Runs and configures the anomalib tiled-ensemble pipeline, which trains/evaluates one model per image tile and merges results (with optional seam smoothing) for high-resolution anomaly detection. Use when the user wants to train with image tiling, mentions "tiled ensemble", or needs to tune tiling/stride/seam-smoothing config. Do not use for regular single-model training (see anomalib-training) or the multi-model benchmarking pipeline (see anomalib-benchmarking).
license
Apache-2.0

Using the Tiled Ensemble Pipeline

The tiled-ensemble pipeline splits each image into overlapping tiles, trains/evaluates a separate model instance per tile position, then merges tile predictions (with optional seam smoothing) back into a full-image anomaly map. Use it for high-resolution images where a single model can't see fine detail at a manageable input size.

Code locations

  • src/anomalib/pipelines/tiled_ensemble/train_pipeline.py — TrainTiledEnsemble: composes the job graph (per-tile training, per-tile prediction, merge, seam smoothing, statistics) and picks SerialRunner or ParallelRunner based on the configured accelerator and available CUDA devices.
  • src/anomalib/pipelines/tiled_ensemble/test_pipeline.py — EvalTiledEnsemble: runs inference/evaluation for an already-trained ensemble.
  • src/anomalib/pipelines/tiled_ensemble/components/ — individual job implementations (model training, prediction, merging, smoothing, metrics).
  • src/anomalib/pipelines/tiled_ensemble/components/utils/ensemble_engine.py — TiledEnsembleEngine, an Engine subclass that customizes per-tile checkpoint/workspace naming.

Running it

bash
python tools/tiled_ensemble/train.py --config tools/tiled_ensemble/ens_config.yaml
python tools/tiled_ensemble/eval.py  --config tools/tiled_ensemble/ens_config.yaml \
                                     --root results/Padim/MVTecAD/bottle/v0

train.py runs TrainTiledEnsemble().run() which includes evaluation after training; eval.py runs EvalTiledEnsemble to re-run evaluation against an existing results directory (--root) — use it only when you want to evaluate again without retraining.

Config structure

Start from tools/tiled_ensemble/ens_config.yaml and adjust the fields you need:

yaml
seed: 42
accelerator: "cuda" # or "cpu"
default_root_dir: "results"

tiling:
  image_size: [256, 256] # size the full image is resized to before tiling
  tile_size: [128, 128] # size of each tile
  stride: 128 # tile stride; stride < tile_size gives overlapping tiles

normalization_stage: image
thresholding_stage: image

data:
  class_path: anomalib.data.MVTecAD
  init_args:
    root: ./datasets/MVTecAD
    category: bottle
    train_batch_size: 32
    eval_batch_size: 32
    num_workers: 8
    val_split_mode: from_test
    test_split_mode: from_dir

SeamSmoothing:
  apply: False
  sigma: 2
  width: 0.1

TrainModels:
  model:
    class_path: Padim

Key fields:

  • tiling.tile_size / tiling.stride — the core tiling geometry; stride < tile_size produces overlap that SeamSmoothing then blends.
  • data.class_path — any image anomalib.data.* datamodule that yields ImageBatch (see anomalib-training / anomalib-adding-a-datamodule). Video and depth datamodules are not supported — the tiled collater uses ImageBatch.collate internally.
  • TrainModels.model.class_path — the model class trained per tile; must be a standard image model that only requires batch.image as input. Models requiring additional inputs (e.g. CFM which needs point_cloud/depth_map) are not compatible with the tiled collater. Video models are also not compatible.
  • SeamSmoothing.apply — when True, applies Gaussian blending at tile boundaries. This is most useful when tiles overlap (stride < tile_size), but can also smooth hard boundaries between non-overlapping tiles. Set False to skip if seam artifacts are not visible.

For a worked reference invocation with a full config, see tests/integration/pipelines/test_tiled_ensemble.py.

Show full SKILL.md (150 more words)Show less

Gotchas

  • The pipeline trains one model instance per tile position, not one shared model — total training cost scales with the number of tiles, not just image count. Budget accordingly before scaling up tile_size/stride combinations.
  • accelerator: cuda with multiple visible GPUs triggers ParallelRunner, which trains multiple tile jobs concurrently across devices — set accelerator: cpu (or restrict visible devices) for deterministic single-process runs while debugging a config.
  • Eval (eval.py) needs --root pointing at the exact output directory produced by the matching training run; it does not re-derive this automatically.
  • data.init_args must include val_split_mode and test_split_mode — the pipeline reads these directly from the config before datamodule defaults are applied, and will raise KeyError if missing.

Reviewer / self-check

  • tiling.tile_size/stride chosen relative to tiling.image_size (stride ≤ tile_size).
  • data.class_path and TrainModels.model.class_path both resolve to real, exported classes.
  • SeamSmoothing.apply is intentional given whether tiles overlap.
  • Training run completed and its results/... path is used correctly as eval.py --root.

© open-edge-platform, Apache-2.0. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

SKILL.md and 1 other file in .agents/skills/anomalib-tiled-ensemble of open-edge-platform/anomalib.

  • SKILL.md
  • evals/evals.json

Open the folder on GitHubat commit dc087d5

Compare with similar skills

Anomalib Tiled Ensemble next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

Anomalib Tiled Ensemble compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Anomalib Tiled Ensemble this skillopen-edge-platform/anomalib6.2k—~1.4kAutomated safety check: PassApache-2.0
Tao Finetune Nv Tesseract Ad DiffusionNVIDIA/skills3.5k—~2.9kAutomated safety check: NotesApache-2.0
TimesFM Forecastinggoogle-research/timesfm34k—~4.7kAutomated safety check: PassApache-2.0
Matlab Engineer Tabular Featuresmatlab/matlab-agentic-toolkit1.1k—~4.8kAutomated safety check: PassCustom licence
Web2 Reconawarexone/Agentic-Bug-Hunter5.3k2 repos~6.4kAutomated safety check: WarnMIT
Scholar Lingjoshzyj/open-scholar-skill168—~6.7kAutomated safety check: PassCustom licence

Similar skills

  • Tao Finetune Nv Tesseract Ad Diffusion

    NVIDIA/skills

    Official

    NV-Tesseract AD Diffusion — diffusion-based anomaly detection and fine-tuning for multivariate time series.

    3.5k GitHub stars~2.9k tokensUpdated yesterday
    Data & AnalyticsAuto-check: notes

  • TimesFM Forecasting

    google-research/timesfm

    Forecasts any univariate time series zero-shot with Google's TimesFM model, returning point forecasts and calibrated prediction intervals without training.

    34k GitHub stars~4.7k tokensUpdated 10 days ago
    Data & AnalyticsAuto-check passed

  • Matlab Engineer Tabular Features

    matlab/matlab-agentic-toolkit

    A skill your agent uses when engineering or selecting the best features for single-response classification or regression in MATLAB, whatever the data's modality — for non-tabular data it routes…

    1.1k GitHub stars~4.8k tokensUpdated yesterday
    Data & AnalyticsAuto-check passed

  • Web2 Recon

    awarexone/Agentic-Bug-Hunter

    Web2 recon pipeline — subdomain enumeration (subfinder, Chaos API, assetfinder), live host discovery (dnsx, httpx), URL crawling (katana, waybackurls, gau), directory fuzzing (ffuf), JS analysis…

    5.3k GitHub starsUsed in 2 repos~6.4k tokens
    SecurityAuto-check: warnings

  • Scholar Ling

    joshzyj/open-scholar-skill

    Design and analyze studies in sociolinguistics, language variation, acoustic phonetics, discourse analysis, language contact, and computational linguistics.

    168 GitHub stars~6.7k tokensUpdated 21 days ago
    Data & AnalyticsAuto-check passed

  • Analyzing API Gateway Access Logs

    mukul975/Anthropic-Cybersecurity-Skills

    Parses API Gateway access logs (AWS API Gateway, Kong, Nginx) to detect BOLA/IDOR attacks, rate limit bypass, credential scanning, and injection attempts.

    34k GitHub stars~581 tokensUpdated 1 mo ago
    Data & AnalyticsAuto-check passed

More from open-edge-platform/anomalib

All 17 skills in this repo

  • Anomalib Adding A Model

    open-edge-platform/anomalib

    Adds a new anomaly-detection model to anomalib under src/anomalib/models/.

    6.2k GitHub stars~1.9k tokensUpdated yesterday
    Auto-check passed

  • Anomalib Benchmarking

    open-edge-platform/anomalib

    Runs the anomalib benchmarking pipeline to train/evaluate a grid of model + dataset (+ category) combinations and collect metrics into a results CSV.

    6.2k GitHub stars~1k tokensUpdated yesterday
    Auto-check passed

  • Anomalib Training

    open-edge-platform/anomalib

    Trains an anomalib model on a dataset via the Python API or CLI, including training on a custom folder-structured dataset with the Folder datamodule.

    6.2k GitHub stars~1.4k tokensUpdated yesterday
    Auto-check passed

  • Model Sample Image Export

    open-edge-platform/anomalib

    Export, validate, and publish model sample-result images into docs/source/images and reference them from README/docs pages.

    6.2k GitHub stars~874 tokensUpdated yesterday
    Auto-check passed

  • UI Test Utils

    open-edge-platform/anomalib

    A skill your agent uses when writing or updating Anomalib Studio UI component or hook tests that need the shared render/renderHook helpers, React Router paths or parameters, React Query, theme…

    6.2k GitHub stars~1.3k tokensUpdated yesterday
    Auto-check passed

  • Anomalib Adding A Datamodule

    open-edge-platform/anomalib

    Adds a new dataset/datamodule to anomalib under src/anomalib/data/.

    6.2k GitHub stars~5.1k tokensUpdated yesterday
    Auto-check passed

Categories

Questions about Anomalib Tiled Ensemble

What does Anomalib Tiled Ensemble do?

Runs and configures the anomalib tiled-ensemble pipeline, which trains/evaluates one model per image tile and merges results (with optional seam smoothing) for high-resolution anomaly detection. Anomalib Tiled Ensemble is an agent skill from open-edge-platform/anomalib. Runs and configures the anomalib tiled-ensemble pipeline, which trains/evaluates one model per image tile and merges results (with optional seam smoothing) for high-resolution anomaly detection.

When should I use Anomalib Tiled Ensemble?

Anomalib Tiled Ensemble fits situations like: the user wants to train with image tiling; mentions tiled ensemble; needs to tune tiling/stride/seam-smoothing config; regular single-model training (see anomalib-training).

How do I install Anomalib Tiled Ensemble in Claude Code?

Run `npx skills add open-edge-platform/anomalib --skill anomalib-tiled-ensemble -a claude-code`. Or copy the skill folder (.agents/skills/anomalib-tiled-ensemble in open-edge-platform/anomalib) into .claude/skills/anomalib-tiled-ensemble in your project. Claude Code loads it when a task matches its description.

How do I install Anomalib Tiled Ensemble in Codex?

Run `npx skills add open-edge-platform/anomalib --skill anomalib-tiled-ensemble -a codex`. Or copy the skill folder (.agents/skills/anomalib-tiled-ensemble in open-edge-platform/anomalib) into .agents/skills/anomalib-tiled-ensemble in your project. Codex loads it when a task matches its description.

Can I use Anomalib Tiled Ensemble in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add open-edge-platform/anomalib --skill anomalib-tiled-ensemble -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/anomalib-tiled-ensemble, .gemini/skills/anomalib-tiled-ensemble, .github/skills/anomalib-tiled-ensemble and .opencode/skills/anomalib-tiled-ensemble in your project.

What does Anomalib Tiled Ensemble need to run?

Going by SKILL.md and its folder, Anomalib Tiled Ensemble needs the command-line tools its instructions call (python). Our summary lists: Python 3.

Does Anomalib Tiled Ensemble access the network?

SKILL.md contains no URLs. Any network use would come from the scripts or tools the agent runs. This is read from the text; nothing was executed.

Is Anomalib Tiled Ensemble safe to install?

Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.

What licence does Anomalib Tiled Ensemble use?

Anomalib Tiled Ensemble is published under the Apache-2.0 licence (declared in SKILL.md). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Anomalib Tiled Ensemble use?

About 1.4k tokens (SKILL.md is roughly 5.4k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.

What are the alternatives to Anomalib Tiled Ensemble?

Skills that share tags, products or a category with Anomalib Tiled Ensemble: Tao Finetune Nv Tesseract Ad Diffusion (NVIDIA/skills, 3.5k stars), TimesFM Forecasting (google-research/timesfm, 34k stars), Matlab Engineer Tabular Features (matlab/matlab-agentic-toolkit, 1.1k stars) and Web2 Recon (awarexone/Agentic-Bug-Hunter, 5.3k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Anomalib Tiled Ensemble?

open-edge-platform (a GitHub organization) maintains it in open-edge-platform/anomalib, which has 6,230 GitHub stars. The repository holds 17 skills in this directory. The repository was last updated on October 8, 2026.

Source: open-edge-platform/anomalib on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.