Update README.md

README.md

@@ -38,7 +38,7 @@ base_model:
 * [Acknowledgments](#acknowledgments)
 * [Citation](#citation)
 
----
+
 
 ## Highlights
 
@@ -47,7 +47,7 @@ base_model:
 * **Same model, two samplers:** Use the displacement sampler for 1-step (or few-step) inference; fall back to Euler along the instantaneous field if you prefer multi-step.
 * **Competitive & fast:** strong ESTOI / SI-SDR / DNSMOS with **very low RTF** on VoiceBank-DEMAND.
 
----
+
 
 ## What’s inside
 
@@ -56,7 +56,7 @@ base_model:
 * **Audio front-end**: complex STFT pipeline; configurable transforms & normalization.
 * **Metrics**: PESQ, ESTOI, SI-SDR; end-to-end **RTF** measurement.
 
----
+
 
 ## Quick start
 
@@ -149,7 +149,7 @@ python evaluate.py \
 > `evaluate.py` writes **enhanced WAVs**.
 > If `--odesolver` is not given, it **auto-picks** (`euler_mf` when MF-SE was used; otherwise `euler`).
 
----
+
 
 ## Configuration
 
@@ -169,7 +169,7 @@ Common flags you may want to tweak:
 
   * Defined in `backbones/` and `SpecsDataModule` (STFT, transforms, normalization)
 
----
+
 
 ## Repository structure
 
@@ -211,19 +211,19 @@ Many design choices (complex STFT pipeline, training infrastructure) are inspire
 
 * **VoiceBank–DEMAND (16 kHz)**: We have hosted the weight files on Google Drive and added the link here.— [Google Drive Link](https://drive.google.com/file/d/1QAxgd5BWrxiNi0q2qD3n1Xcv6bW0X86-/view?usp=sharing)
 
----
+
 
 ## Acknowledgments
 
 We gratefully acknowledge **Prof. Xie Chen’s group (X-LANCE Lab, SJTU)** for their **valuable guidance and support** on training practices and engineering tips that helped this work a lot.
 
----
+
 
 ## Citation
 
 * **Citation:** The paper is currently under review. We will add a BibTeX entry and article link once available.
 
----
+
 
 
 **Questions or issues?** Please open a GitHub issue or pull request.