Update README.md

README.md

@@ -2,19 +2,30 @@
 license: apache-2.0
 ---
 
-This model is meant to be used on speechbrain.
+## Demo
 
-The model is implemented on a fork of speechbrain, with the yaml written for the WSJ0-2mix dataset.
-https://github.com/Yip-Jia-Qi/speechbrain/tree/add_spgm/recipes/WSJ0Mix/separation
+A demo with instructions on how to run inference on the model is available as a colab notebook [here](https://colab.research.google.com/drive/1zKEaRFNITve7WPsqVNUuaRXiduR7H1Ki?usp=sharing)
 
-Alternatievly, I have created a standalone version of this model with minimal dependencies here:
-https://github.com/Yip-Jia-Qi/spgm_standalone
+The standalone version of this model for inference with minimal dependencies [here](https://github.com/Yip-Jia-Qi/spgm_standalone)
 
+Training is handled by speechbrain. This can be done through my fork of the speechbrain repository found [here](https://github.com/Yip-Jia-Qi/speechbrain/tree/add_spgm).
 
-| | SPGM, WSJ0-2Mix |
-|--- | --- |
-|SpeedPerturb | 22.1 |
-|DynamicMixing | 22.7 |
+## Results
+Here are the SI - SNRi results (in dB) on the test set of WSJ0-2 Mix:
+
+|Model| Data Augmentation | WSJ0-2Mix (SI-SNRi)|
+| --- |--- | --- |
+|spgm (paper)|SpeedPerturb | 22.1 |
+|[spgm-base](https://huggingface.co/yipjiaqi/spgm-base)|DynamicMixing | 22.7 |
+|[spgm-opt](https://huggingface.co/yipjiaqi/spgm-opt)|DynamicMixing | 23.0 |
+
+In the original paper accepted to ICASSP, the only data augmentation used was speed perturbation. Subsequently we trained the model using dynamic mixing, which yielded improvements in performance. 
+
+Additionally, after further exploring some hyperparameters, we obtain an optimized version of SPGM, spgm-opt that achieved 23.0dB SI-SDRi
+
+The weights and config of spgm-base and spgm-opt have been uploaded to huggingface and can be accessed using the code in the spgm_standalone [repo](https://github.com/Yip-Jia-Qi/spgm_standalone).
+
+## Citation
 
 If you find this model useful, please cite:
 ```bibtex