Credit training dataset (DDD-Cambodia/khmer-speech-dataset)

README.md

@@ -4,6 +4,8 @@ language:
 - km
 base_model:
 - Qwen/Qwen3-ASR-0.6B
+datasets:
+- DDD-Cambodia/khmer-speech-dataset
 pipeline_tag: automatic-speech-recognition
 library_name: transformers
 tags:
@@ -43,8 +45,10 @@ model-index:
 
 A Khmer (ខ្មែរ) automatic speech recognition model, fine-tuned from
 [**Qwen/Qwen3-ASR-0.6B**](https://huggingface.co/Qwen/Qwen3-ASR-0.6B) on
-~700 hours of Khmer speech. It substantially improves Khmer transcription
-accuracy over the base model while keeping the compact 0.6B footprint.
+~700 hours of Khmer speech from the
+[**DDD-Cambodia/khmer-speech-dataset**](https://huggingface.co/datasets/DDD-Cambodia/khmer-speech-dataset).
+It substantially improves Khmer transcription accuracy over the base model
+while keeping the compact 0.6B footprint.
 
 ## Results
 
@@ -101,7 +105,7 @@ chunked; for long clips set a larger `max_new_tokens` so the transcript is not c
 |---|---|
 | Base model | Qwen/Qwen3-ASR-0.6B |
 | Language | Khmer (`km`) |
-| Training data | ~700 h Khmer speech (~384k clips) |
+| Training data | [DDD-Cambodia/khmer-speech-dataset](https://huggingface.co/datasets/DDD-Cambodia/khmer-speech-dataset) (~700 h, ~384k clips) |
 | Epochs | 3 (35,997 steps) |
 | Effective batch size | 32 (per-device 4 × grad-accum 8) |
 | Learning rate | 2e-5, linear schedule with warmup |