Update README.md

README.md

@@ -8,5 +8,11 @@ dataset_info:
     - name: source
       dtype: string
 ---
-# Vietnamese Audio Filtered Dataset
-Filtered Vietnamese audio dataset with ASR-verified transcripts.
+# Vivoice relabeled using Qwen3-ASR
+
+The original dataset: [capleaf/viVoice](https://huggingface.co/datasets/capleaf/viVoice)
+Sample rate: 44.1 kHz
+
+> [!NOTE]  
+> **How-tos**: I use the [Qwen/Qwen3-ASR-1.7B](https://huggingface.co/Qwen/Qwen3-ASR-1.7B) models then update the label for the sample (audio - text) which have WER (word error rate) < 15%.
+