Update README.md
Browse files
README.md
CHANGED
|
@@ -85,3 +85,14 @@ while let samples = getAudioChunk() {
|
|
| 85 |
## Source
|
| 86 |
|
| 87 |
Original model: [nvidia/diar_streaming_sortformer_4spk-v2.1](https://huggingface.co/nvidia/diar_streaming_sortformer_4spk-v2.1)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 85 |
## Source
|
| 86 |
|
| 87 |
Original model: [nvidia/diar_streaming_sortformer_4spk-v2.1](https://huggingface.co/nvidia/diar_streaming_sortformer_4spk-v2.1)
|
| 88 |
+
|
| 89 |
+
|
| 90 |
+
|
| 91 |
+
## Credits & Acknowledgements
|
| 92 |
+
|
| 93 |
+
Special thanks to **[@grd]** for their significant contributions to this project, including:
|
| 94 |
+
* **CoreML Conversion:** Architecting the `export_gradient_descent.py` scripts and handling the static shape constraints for Apple Silicon.
|
| 95 |
+
* **Model Optimization:** Fine-tuning the Pre-Encoder and Head split to ensure compatibility with the Apple Neural Engine (ANE).
|
| 96 |
+
* **Research & Debugging:** Porting the complex streaming state logic from NeMo to CoreML-compatible PyTorch wrappers.
|
| 97 |
+
|
| 98 |
+
This project was built upon the foundational work of the NVIDIA NeMo team.
|