Datasets:
Upload README.md with huggingface_hub
Browse files
README.md
CHANGED
|
@@ -1,45 +1,68 @@
|
|
| 1 |
---
|
| 2 |
license: odc-by
|
| 3 |
-
|
| 4 |
-
|
| 5 |
-
|
| 6 |
-
-
|
| 7 |
-
|
| 8 |
-
|
| 9 |
-
|
| 10 |
-
|
| 11 |
-
|
| 12 |
-
-
|
| 13 |
-
dtype: string
|
| 14 |
-
splits:
|
| 15 |
-
- name: validation
|
| 16 |
-
num_bytes: 235869
|
| 17 |
-
num_examples: 481
|
| 18 |
-
download_size: 113320
|
| 19 |
-
dataset_size: 235869
|
| 20 |
-
configs:
|
| 21 |
-
- config_name: default
|
| 22 |
-
data_files:
|
| 23 |
-
- split: validation
|
| 24 |
-
path: data/validation-*
|
| 25 |
---
|
| 26 |
|
| 27 |
-
#
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 28 |
|
| 29 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 30 |
|
| 31 |
## Citation
|
|
|
|
| 32 |
```bibtex
|
| 33 |
-
@misc{
|
| 34 |
-
title = {sst2-ka:
|
| 35 |
author = {Tbilisi AI Lab Team},
|
| 36 |
year = {2025},
|
| 37 |
publisher = {Hugging Face},
|
| 38 |
-
howpublished = {\url{https://huggingface.co/datasets/tbilisi-ai-lab/sst2-ka}}
|
| 39 |
-
note = {Accessed: 2025-10-18}
|
| 40 |
}
|
| 41 |
```
|
| 42 |
|
| 43 |
-
## License
|
| 44 |
-
|
|
|
|
|
|
|
|
|
|
| 45 |
|
|
|
|
|
|
|
|
|
| 1 |
---
|
| 2 |
license: odc-by
|
| 3 |
+
task_categories:
|
| 4 |
+
- text-classification
|
| 5 |
+
language:
|
| 6 |
+
- ka
|
| 7 |
+
- en
|
| 8 |
+
size_categories:
|
| 9 |
+
- n<1K
|
| 10 |
+
tags:
|
| 11 |
+
- benchmark
|
| 12 |
+
- sentiment-analysis
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 13 |
---
|
| 14 |
|
| 15 |
+
# sst2-ka
|
| 16 |
+
|
| 17 |
+
Georgian translation of the SST-2 (Stanford Sentiment Treebank) benchmark.
|
| 18 |
+
|
| 19 |
+
## Dataset Summary
|
| 20 |
+
|
| 21 |
+
| Property | Value |
|
| 22 |
+
|----------|-------|
|
| 23 |
+
| Examples | 481 |
|
| 24 |
+
| Splits | validation |
|
| 25 |
+
| Languages | Georgian, English |
|
| 26 |
+
| Task | Sentiment Classification |
|
| 27 |
+
|
| 28 |
+
## Data Fields
|
| 29 |
+
|
| 30 |
+
- `sentence`: Sentence (English)
|
| 31 |
+
- `label`: Sentiment label (0=negative, 1=positive)
|
| 32 |
+
- `idx`: Example index
|
| 33 |
+
- `sentence_ka`: Sentence (Georgian)
|
| 34 |
+
|
| 35 |
+
## Translation Methodology
|
| 36 |
|
| 37 |
+
1. Translation model generates initial Georgian translation
|
| 38 |
+
2. Human contractor verifies and fixes translation if needed
|
| 39 |
+
3. Human validator confirms translation quality
|
| 40 |
+
|
| 41 |
+
## Usage
|
| 42 |
+
|
| 43 |
+
```python
|
| 44 |
+
from datasets import load_dataset
|
| 45 |
+
|
| 46 |
+
ds = load_dataset("tbilisi-ai-lab/sst2-ka")
|
| 47 |
+
```
|
| 48 |
|
| 49 |
## Citation
|
| 50 |
+
|
| 51 |
```bibtex
|
| 52 |
+
@misc{tbilisi2025sst2ka,
|
| 53 |
+
title = {sst2-ka: Georgian SST-2 Benchmark},
|
| 54 |
author = {Tbilisi AI Lab Team},
|
| 55 |
year = {2025},
|
| 56 |
publisher = {Hugging Face},
|
| 57 |
+
howpublished = {\url{https://huggingface.co/datasets/tbilisi-ai-lab/sst2-ka}}
|
|
|
|
| 58 |
}
|
| 59 |
```
|
| 60 |
|
| 61 |
+
## License
|
| 62 |
+
|
| 63 |
+
[ODC-BY](https://opendatacommons.org/licenses/by/1-0/)
|
| 64 |
+
|
| 65 |
+
## Contact
|
| 66 |
|
| 67 |
+
- **Organization:** [Tbilisi AI Lab](https://huggingface.co/tbilisi-ai-lab)
|
| 68 |
+
- **Website:** [ailab.ge](https://ailab.ge)
|