# FAIO Warmup Competition 1 (2025) Long date version — Data

来源：https://www.kaggle.com/competitions/faio-warmup-competition-1-2025-long-date-version/data

采集时间：2026-10-19T08:47:12Z

## Dataset Description

The dataset contains 3D point cloud representations of geometric shapes. Each shape is expressed as a list of (x, y, z) coordinates, flattened into a single row. Your task is to classify each shape based on its structure.

Each file is described below:

- **`train.csv`** — each row corresponds to a labeled shape. The values are grouped in triplets (x, y, z), and the **last number** is the shape’s class label (e.g., 0 = sphere, 1 = cube, etc.).
  
  Example row:
0.12,0.94,0.32, 0.15,0.89,0.35, ..., 0.42,0.70,0.18, 2

This row describes a shape composed of multiple 3D points, and it belongs to class `2`.

- **`test.csv`** — same format as `train.csv`, but **without the class label** at the end. Your task is to predict the class label for each shape in this file.

Example row:
0.28,0.64,0.10, 0.36,0.61,0.12, ..., 0.45,0.55,0.09

- **`sample_submission.csv`** — shows the expected submission format:
id,label
0,1
1,3
2,0
...

Where `id` is the row number in `test.csv` and `label` is your predicted class (integer).

### Notes

- The number of points per shape may differ across datasets.
- All coordinate values are normalized to the range [0, 1].
- The number of unique classes can be inferred from the `train.csv` file.