README.md

---
tags:
- smart-manufacturing
- sft
- industrial
- vision
license: other
pretty_name: "193"
extra_gated_fields:
  Name: text
  Affiliation: text
  Intended use: text
extra_gated_prompt: >-
  This dataset is released for **research use**. Access is reviewed and granted
  **manually** by the maintainers. Please state your name, affiliation, and intended use.
---

# 193

Steel-sheet surface anomaly detection (binary; 4 anonymous class ids + segmentation kept as GT). Category **B**, task **T-B1**, in the unified Smart-Manufacturing SFT schema.

> The repository name is an internal task code. See **Provenance** below for the underlying dataset.

## Records

**12,568** records (train=12568). Pixel masks are embedded as a `mask` image column.

## Unified SFT schema

| field | type | meaning |
|---|---|---|
| `query` | str | the question / instruction (model input) |
| `image` | Image | the input image (bytes embedded) |
| `annot` | str | the answer — for this dataset: the plain-text image-level label `good` or `anomalous`. Severstal's 4 official defect classes are UNNAMED (numeric ids only), so — following the same principle as DAGM — the anonymous class id is NOT asked of the model; the per-image class-id list and the RLE-decoded class-indexed segmentation `mask` are kept as deferred ground truth in `metadata.defect_classes` and the `mask` column — see **Task, mask & split** below |
| `reasoning` | null | no native CoT in these datasets |
| `cate` | "B" | SFT category |
| `task` | "T-xx" | unified task id |
| `metadata` | str (JSON) | split, provenance, `image_path`, `image_sha256` (dedup key) |
| `mask` | Image \| null | *(T-B1/T-B2 only)* the pixel ground-truth mask, bytes embedded |
| `masks` | list[Image] | *(D21 only)* multi-region masks |

## Task, mask & split

**What this is.** Severstal Steel Defect Detection (Kaggle 2019) — originally pixel-level **4-class defect
segmentation** of steel-sheet surfaces (1600×256). The four defect classes are official but **UNNAMED**:
Severstal (the steel maker) never released what they physically mean — the labels are only the numeric ids
**1–4**. Data is obtained from the public HF mirror `rohanath/severstal-steel-detection`.

**Query & answer — why BINARY (this repo's SFT task).** Because the class ids are anonymous, there is no
semantic concept for a vision-language model to ground: asking it to output `1` vs `3` would be asking it to
reproduce an arbitrary, meaningless label. So — exactly as we do for **DAGM** (also anonymous classes) — the
task is **binary**: `query` asks only good vs anomalous, and `annot` is the plain-text label `good` or
`anomalous` (label = `anomalous` iff the image has a defect annotation). **We do not put the anonymous class
id in `annot`.**

**Anonymous class ids + mask (deferred GT, in metadata / mask column).** The class information is not
discarded — it is preserved as ground truth: `metadata.defect_classes` holds the per-image list of defect
class ids present (e.g. `[1, 3]`; 427 images carry more than one), and the competition's run-length-encoded
(RLE) masks are decoded (column-major) into a **class-indexed** segmentation `mask` (pixel value = defect
class id), kept as deferred localization GT with `defect_area_fraction` in `metadata`. A downstream user who
obtains a semantic naming for the 4 classes can recover the full multi-label / segmentation task from these.

**What the 4 classes are — UNOFFICIAL, community observation (pending manual verification).** Severstal never
released the meaning of classes 1–4; the numeric ids remain the only official labels and `annot` stays binary.
For reference only, independent community exploratory analyses of the competition (Kaggle discussions and
public write-ups) consistently describe the classes by *appearance* — **not** official names — roughly as:
Class **1** = small localized spots / pinhole-like marks; Class **2** = thin small linear marks (a small
crack-like defect); Class **3** = larger linear defects / scratches with distinct edges (classes 2 and 3 look
similar and are hard to tell apart); Class **4** = large surface patches (the most visually distinct). The
class frequencies in this dataset — **3 ≫ 1 > 4 > 2** (5,150 / 897 / 801 / 247) — match those write-ups.
**These are unofficial community descriptions, NOT Severstal's own defect taxonomy — they are provided only to
help interpret `metadata.defect_classes` and are subject to manual verification; do not treat them as
ground-truth type names.** (A "pitted / crazing / scratches / patches" naming also circulates online but
appears to be borrowed from the unrelated NEU-DET dataset, so it is not adopted here.)

**What is excluded (audit).** The Kaggle **test set** (5,506 images, GT withheld) is dropped; the mirror's
`visualized_images/` are mask **overlays** (would leak the answer) and are excluded; the mirror's
`labels/*.txt` are **derived YOLO boxes** (computed from the RLE) — not used, the RLE is the raw GT.

**Split.** Single `train` pool = **12,568** images: 6,666 defective (RLE-annotated) + 5,902 defect-free.
The mirror's degenerate train(all-defective)/val(all-good) partition is not used as a split.

## Provenance

Underlying dataset: **Severstal**. Upstream license: **other (Kaggle competition data; via HF mirror rohanath/severstal-steel-detection)** (this card is `license: other`; respect the upstream terms). Converted read-only from the raw source into the unified schema; conversion script: `193/convert_d93.py`, published with `publish/push_to_hf.py`, both in [`AI4Manufacturing/forge_model`](https://github.com/AI4Manufacturing/forge_model).

## Overlap / de-duplication (§8)

The Kaggle test set (GT withheld) and the mirror's derived YOLO labels are excluded; the RLE is the raw GT. Each record carries `metadata.image_sha256` so overlapping images can be kept entirely on one side of a train/eval split.