upd md

.gitignore

@@ -0,0 +1 @@
+rename.sh

README.md

@@ -1,3 +1,31 @@
 ---
 license: mit
+language:
+- en
+metrics:
+- accuracy
+tags:
+- music
+- art
+library_name: timm
 ---
+
+# 简介 Intro
+字节孔秋强高分辨率钢琴转录系统是一款先进的音乐信息检索工具，专门设计用于高精度地自动转换钢琴演奏的音频信号为详细的乐谱。该系统的训练模型镜像采用了最新的深度学习技术，包括卷积神经网络和循环神经网络，以实现对音符时序和音高的精确捕捉。通过多尺度特征学习和长期依赖关系的建模，该系统能够处理复杂的音乐结构，并提供高密度音符的准确转录。字节孔秋强系统不仅优化了音乐分析和理论研究的效率，也为音乐教育和表演提供了强大的技术支持。
+
+Qiuqiang Kong High-Resolution Piano Transcription System of ByteDance is an advanced music information retrieval tool specifically designed to automatically convert audio signals from piano performances into detailed sheet music with high accuracy. The system's training model mirrors the latest deep learning techniques, including convolutional neural networks and recurrent neural networks, to achieve accurate capture of note timing and pitch. Through multi-scale feature learning and modeling of long-term dependencies, the system is able to handle complex musical structures and provide accurate transcription of high-density notes. This ByteDance system not only optimizes the efficiency of music analysis and theoretical research, but also provides powerful technical support for music education and performance.
+
+## 快速使用 Usage
+```python
+from modelscope import snapshot_download
+model_dir = snapshot_download("MuGeminorum/piano_transcription")
+```
+
+## 维护 Maintenance
+```bash
+git clone git@hf.co:MuGeminorum/piano_transcription
+cd piano_transcription
+```
+
+## 参考引用 Reference
+[1] [High-resolution Piano Transcription with Pedals by Regressing Onset and Offset Times](https://arxiv.org/pdf/2010.01815)<br>