Datasets:
update README
Browse files- README.md +31 -27
- README_ZH.md +32 -28
README.md
CHANGED
|
@@ -58,8 +58,38 @@ SongFormBench is a **carefully curated, expert-annotated benchmark** designed to
|
|
| 58 |
|
| 59 |
---
|
| 60 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 61 |
## 🎼 Mel Spectrogram Details
|
| 62 |
|
|
|
|
|
|
|
|
|
|
| 63 |
Environment configuration can refer to the official implementation of BigVGan. If the audio source becomes invalid, you can reconstruct the audio using the following method.
|
| 64 |
|
| 65 |
### 🎸 SongFormBench-HarmonixSet
|
|
@@ -84,33 +114,7 @@ Reproduce using [**bigvgan_v2_44khz_128band_256x**](https://huggingface.co/nvidi
|
|
| 84 |
# See implementation
|
| 85 |
utils/CN/infer.py
|
| 86 |
```
|
| 87 |
-
|
| 88 |
-
---
|
| 89 |
-
|
| 90 |
-
## 🚀 Getting Started
|
| 91 |
-
|
| 92 |
-
### Quick Load
|
| 93 |
-
```python
|
| 94 |
-
from datasets import load_dataset
|
| 95 |
-
|
| 96 |
-
# Load the complete benchmark
|
| 97 |
-
dataset = load_dataset("ASLP-lab/SongFormBench")
|
| 98 |
-
```
|
| 99 |
-
|
| 100 |
-
---
|
| 101 |
-
|
| 102 |
-
## 📚 Resources & Links
|
| 103 |
-
|
| 104 |
-
- 📖 Paper: *coming soon*
|
| 105 |
-
- 💻 Code: [GitHub Repository](https://github.com/ASLP-lab/SongFormer)
|
| 106 |
-
- 🧑💻 Model: [SongFormer](https://huggingface.co/ASLP-lab/SongFormer)
|
| 107 |
-
- 📂 Dataset: [SongFormDB](https://huggingface.co/datasets/ASLP-lab/SongFormDB)
|
| 108 |
-
|
| 109 |
-
---
|
| 110 |
-
|
| 111 |
-
## 🤝 Citation
|
| 112 |
-
|
| 113 |
-
comming soon.
|
| 114 |
|
| 115 |
---
|
| 116 |
|
|
|
|
| 58 |
|
| 59 |
---
|
| 60 |
|
| 61 |
+
## 🚀 Getting Started
|
| 62 |
+
|
| 63 |
+
### Quick Load
|
| 64 |
+
```python
|
| 65 |
+
from datasets import load_dataset
|
| 66 |
+
|
| 67 |
+
# Load the complete benchmark
|
| 68 |
+
dataset = load_dataset("ASLP-lab/SongFormBench")
|
| 69 |
+
```
|
| 70 |
+
|
| 71 |
+
---
|
| 72 |
+
|
| 73 |
+
## 📚 Resources & Links
|
| 74 |
+
|
| 75 |
+
- 📖 Paper: *coming soon*
|
| 76 |
+
- 💻 Code: [GitHub Repository](https://github.com/ASLP-lab/SongFormer)
|
| 77 |
+
- 🧑💻 Model: [SongFormer](https://huggingface.co/ASLP-lab/SongFormer)
|
| 78 |
+
- 📂 Dataset: [SongFormDB](https://huggingface.co/datasets/ASLP-lab/SongFormDB)
|
| 79 |
+
|
| 80 |
+
---
|
| 81 |
+
|
| 82 |
+
## 🤝 Citation
|
| 83 |
+
|
| 84 |
+
comming soon.
|
| 85 |
+
|
| 86 |
+
---
|
| 87 |
+
|
| 88 |
## 🎼 Mel Spectrogram Details
|
| 89 |
|
| 90 |
+
<details>
|
| 91 |
+
<summary>Click to expand/collapse</summary>
|
| 92 |
+
|
| 93 |
Environment configuration can refer to the official implementation of BigVGan. If the audio source becomes invalid, you can reconstruct the audio using the following method.
|
| 94 |
|
| 95 |
### 🎸 SongFormBench-HarmonixSet
|
|
|
|
| 114 |
# See implementation
|
| 115 |
utils/CN/infer.py
|
| 116 |
```
|
| 117 |
+
</details>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 118 |
|
| 119 |
---
|
| 120 |
|
README_ZH.md
CHANGED
|
@@ -46,8 +46,39 @@ SongFormBench 是一个**经过精心整理、由专家标注的基准数据集*
|
|
| 46 |
|
| 47 |
---
|
| 48 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 49 |
## 🎼 梅尔频谱图细节
|
| 50 |
|
|
|
|
|
|
|
|
|
|
| 51 |
环境配置可参考 BigVGan 的官方实现。如果音频源失效，可使用以下方法重建音频。
|
| 52 |
|
| 53 |
### 🎸 SongFormBench-HarmonixSet
|
|
@@ -72,35 +103,8 @@ bash inference_e2e.sh
|
|
| 72 |
# 查看实现
|
| 73 |
utils/CN/infer.py
|
| 74 |
```
|
|
|
|
| 75 |
|
| 76 |
-
---
|
| 77 |
-
|
| 78 |
-
## 🚀 快速开始
|
| 79 |
-
|
| 80 |
-
### 快速加载
|
| 81 |
-
```python
|
| 82 |
-
from datasets import load_dataset
|
| 83 |
-
|
| 84 |
-
# 加载完整基准数据集
|
| 85 |
-
dataset = load_dataset("ASLP-lab/SongFormBench")
|
| 86 |
-
```
|
| 87 |
-
|
| 88 |
-
---
|
| 89 |
-
|
| 90 |
-
## 📚 资源与链接
|
| 91 |
-
|
| 92 |
-
- 📖 论文：*即将发布*
|
| 93 |
-
- 💻 代码：[GitHub 仓库](https://github.com/ASLP-lab/SongFormer)
|
| 94 |
-
- 🧑💻 模型：[SongFormer](https://huggingface.co/ASLP-lab/SongFormer)
|
| 95 |
-
- 📂 数据集：[SongFormDB](https://huggingface.co/datasets/ASLP-lab/SongFormDB)
|
| 96 |
-
|
| 97 |
-
---
|
| 98 |
-
|
| 99 |
-
## 🤝 引用
|
| 100 |
-
|
| 101 |
-
即将发布。
|
| 102 |
-
|
| 103 |
-
---
|
| 104 |
|
| 105 |
## 📧 联系方式
|
| 106 |
|
|
|
|
| 46 |
|
| 47 |
---
|
| 48 |
|
| 49 |
+
## 🚀 快速开始
|
| 50 |
+
|
| 51 |
+
### 快速加载
|
| 52 |
+
```python
|
| 53 |
+
from datasets import load_dataset
|
| 54 |
+
|
| 55 |
+
# 加载完整基准数据集
|
| 56 |
+
dataset = load_dataset("ASLP-lab/SongFormBench")
|
| 57 |
+
```
|
| 58 |
+
|
| 59 |
+
---
|
| 60 |
+
|
| 61 |
+
## 📚 资源与链接
|
| 62 |
+
|
| 63 |
+
- 📖 论文：*即将发布*
|
| 64 |
+
- 💻 代码：[GitHub 仓库](https://github.com/ASLP-lab/SongFormer)
|
| 65 |
+
- 🧑💻 模型：[SongFormer](https://huggingface.co/ASLP-lab/SongFormer)
|
| 66 |
+
- 📂 数据集：[SongFormDB](https://huggingface.co/datasets/ASLP-lab/SongFormDB)
|
| 67 |
+
|
| 68 |
+
---
|
| 69 |
+
|
| 70 |
+
## 🤝 引用
|
| 71 |
+
|
| 72 |
+
即将发布。
|
| 73 |
+
|
| 74 |
+
---
|
| 75 |
+
|
| 76 |
+
|
| 77 |
## 🎼 梅尔频谱图细节
|
| 78 |
|
| 79 |
+
<details>
|
| 80 |
+
<summary>Click to expand/collapse</summary>
|
| 81 |
+
|
| 82 |
环境配置可参考 BigVGan 的官方实现。如果音频源失效，可使用以下方法重建音频。
|
| 83 |
|
| 84 |
### 🎸 SongFormBench-HarmonixSet
|
|
|
|
| 103 |
# 查看实现
|
| 104 |
utils/CN/infer.py
|
| 105 |
```
|
| 106 |
+
</details>
|
| 107 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 108 |
|
| 109 |
## 📧 联系方式
|
| 110 |
|