Update README.md

README.md

@@ -1,4 +1,9 @@
 ---
+license: cc-by-nc-sa-4.0
+task_categories: ["automatic-speech-recognition", "text-to-speech"]
+language: ["vi"]
+pretty_name: FPT Open Speech Dataset (FOSD)
+size_categories: ["10K<n<100K"]
 dataset_info:
   features:
   - name: audio
@@ -17,3 +22,34 @@ configs:
   - split: train
     path: data/train-*
 ---
+
+# unofficial mirror of FPT Open Speech Dataset (FOSD)
+
+released publicly in 2018 by FPT Corporation
+
+100h, 26k samples
+
+official link (dead): https://fpt.ai/fpt-open-speech-data/
+
+mirror: https://data.mendeley.com/datasets/k9sxg2twv4/4
+
+DOI: `10.17632/k9sxg2twv4.4`
+
+pre-process:
+- remove non-sense strings: `-N` `\r\n`
+- remove 4 files because missing transcription:
+  - `Set001_V0.1_008210.mp3`
+  - `Set001_V0.1_010753.mp3`
+  - `Set001_V0.1_011477.mp3`
+  - `Set001_V0.1_011841.mp3`
+
+usage with HuggingFace:
+```python
+# pip install -q "datasets[audio]"
+from datasets import load_dataset
+from torch.utils.data import DataLoader
+
+dataset = load_dataset("doof-ferb/fpt_fosd", split="train", streaming=True)
+dataset.set_format(type="torch", columns=["audio", "transcription"])
+dataloader = DataLoader(dataset, batch_size=4)
+```