README.md

---
license: cc-by-nc-4.0
task_categories:
- text-classification
task_ids:
- multi-class-classification
language:
- te
multilinguality:
- monolingual
size_categories:
- n<1K
pretty_name: Padyam2GadyamMeterClassification
tags:
- poetry
- telugu
- meter-classification
- chandassu
- text-classification
- mteb
- poetrymteb
- embedding-evaluation
annotations_creators:
- expert-generated
source_datasets:
- TeluguLLMResearch/Padyam2Gadyam
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
  - split: test
    path: data/test-*
  default: true
dataset_info:
- config_name: default
  features:
  - name: id
    dtype: string
  - name: poem
    dtype: string
  - name: label
    dtype: int64
  - name: label_name
    dtype: string
  splits:
  - name: train
    num_examples: 468
  - name: test
    num_examples: 118
---

# Padyam2GadyamMeterClassification

Single-label **Telugu poetic meter (chandassu) classification** for PoetryMTEB embedding evaluation, derived from [Padyam2Gadyam](https://huggingface.co/datasets/TeluguLLMResearch/Padyam2Gadyam).

## Dataset Card

| Item | Description |
|------|-------------|
| **Source** | [TeluguLLMResearch/Padyam2Gadyam](https://huggingface.co/datasets/TeluguLLMResearch/Padyam2Gadyam); paper [arXiv:2606.02806](https://arxiv.org/abs/2606.02806) |
| **Languages** | Telugu (`te`) |
| **Size** | train=468; test=118 (no validation; MTEB Classification uses train→fit, test→score) |
| **Classes** | 7 meters with frequency ≥ 10 |
| **Filtering** | Dropped 14 poems in rare meters (count &lt; 10); raw=600 → kept=586 |
| **Splits** | Stratified by meter ≈ 80% / 19% (`train` / `test`), seed=42 |
| **License** | [CC BY-NC 4.0](https://creativecommons.org/licenses/by-nc/4.0/) (same as upstream) |
| **Evaluation metrics** | Classification on embeddings: **accuracy**, **macro/weighted F1** |

---

## Features

| Field | Type | Description |
|-------|------|-------------|
| `id` | string | Example id |
| `poem` | string | Classical Telugu poem text |
| `label` | int64 | Meter class index (0 … C−1) |
| `label_name` | string | Meter name in Telugu |

Codebook: `label_taxonomy.json`.

### Label inventory (kept)

| id | label_name | gloss | count |
|---:|------------|-------|------:|
| 0 | కందం | Kandam (坎达体) | 216 |
| 1 | ఉత్పలమాల | Utpalamāla (优钵罗鬘) | 115 |
| 2 | చంపకమాల | Champakamāla (瞻波迦鬘) | 79 |
| 3 | మత్తేభము | Mattebhamu (醉象律) | 79 |
| 4 | శార్దూలము | Śārdūlamu (虎律) | 43 |
| 5 | తేటగీతి | Tēṭagīti (澄歌体) | 35 |
| 6 | ఆటవెలది | Āṭaveladi (戏舞体) | 19 |

### Dropped meters (count &lt; 10)

| meter | count |
|-------|------:|
| సీసము+తేటగీతి | 6 |
| మాలిని | 2 |
| చంపకమాల-పంచపాది | 2 |
| తరళము | 1 |
| ఉత్పలమాల-పంచపాది | 1 |
| ఉత్సాహము | 1 |
| మత్తకోకిల | 1 |

---

## Construction method

1. Load [TeluguLLMResearch/Padyam2Gadyam](https://huggingface.co/datasets/TeluguLLMResearch/Padyam2Gadyam).
2. Count `Meter`; keep classes with frequency ≥ 10.
3. Map kept meters to contiguous `label` ids (frequency-descending order).
4. Stratified train/test split by meter (no validation split).

---

## How to load

```python
from datasets import load_dataset

ds = load_dataset("PoetryMTEB/Padyam2GadyamMeterClassification")
print(ds["test"][0]["label_name"], ds["test"][0]["poem"][:60])
```

For embedding evaluation, encode `poem`, fit a classifier on `train` labels, score on `test`.

---

## License

Distributed under **Creative Commons Attribution-NonCommercial 4.0 International (CC BY-NC 4.0)**, consistent with the upstream dataset.

---

## Citation / provenance

```bibtex
@misc{kranti2026translatingclassicalpoetrymodern,
      title={Translating Classical Poetry into Modern Prose}, 
      author={Chalamalasetti Kranti and Sowmya Vajjala},
      year={2026},
      eprint={2606.02806},
      archivePrefix={arXiv},
      primaryClass={cs.CL},
      url={https://arxiv.org/abs/2606.02806}, 
}
```

- Dataset: [https://huggingface.co/datasets/TeluguLLMResearch/Padyam2Gadyam](https://huggingface.co/datasets/TeluguLLMResearch/Padyam2Gadyam)
- Paper: [https://arxiv.org/abs/2606.02806](https://arxiv.org/abs/2606.02806)
- This Hub packaging: `PoetryMTEB/Padyam2GadyamMeterClassification`