docs: Add successor banner pointing to SR-CorrNet-SS (V2)

Direct users to the SR-CorrNet-SS repository, the V2 of SepReformer,
which ships all pretrained checkpoints on Hugging Face Hub and a
clean Python library API. Apologize for unresponded issues here.

Author: dmlguq456

README.md

@@ -1,16 +1,10 @@
 
 # SepReformer for Speech Separation [NeurIPS 2024]
 
+> ⚠️ The successor **[SR-CorrNet-SS](https://github.com/dmlguq456/SR_CorrNet_SS)** (V2, [arXiv:2603.29097](https://arxiv.org/abs/2603.29097)) ships all `.pt` checkpoints on HF Hub and a Python library API — please use it instead. Sorry for the unanswered issues here.
 
 This is the official implementation of “Separate and Reconstruct: Asymmetric Encoder-Decoder for Speech Separation” accepted in NeurIPS 2024 [Paper Link(Arxiv)](https://arxiv.org/abs/2406.05983)
 
-[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/separate-and-reconstruct-asymmetric-encoder/speech-separation-on-wsj0-2mix)](https://paperswithcode.com/sota/speech-separation-on-wsj0-2mix?p=separate-and-reconstruct-asymmetric-encoder)
-[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/separate-and-reconstruct-asymmetric-encoder/speech-separation-on-wham)](https://paperswithcode.com/sota/speech-separation-on-wham?p=separate-and-reconstruct-asymmetric-encoder)
-[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/separate-and-reconstruct-asymmetric-encoder/speech-separation-on-whamr)](https://paperswithcode.com/sota/speech-separation-on-whamr?p=separate-and-reconstruct-asymmetric-encoder)
-
-
-
-
 ## News
 
 🔥 October, 2024: We have added feature for single audio inference on our SepReformer-B for WSJ0-2MIX in `models/SepReformer_Base_WSJ0`. You can directly inference by your own samples!