Commit
·
8779c37
1
Parent(s):
5973fe6
Update README
Browse files
README.md
CHANGED
|
@@ -9,9 +9,9 @@ size_categories:
|
|
| 9 |
- 10K<n<100K
|
| 10 |
---
|
| 11 |
|
| 12 |
-
|
| 13 |
|
| 14 |
-
|
| 15 |
<table>
|
| 16 |
<thead>
|
| 17 |
<td>画像</td>
|
|
@@ -29,7 +29,7 @@ size_categories:
|
|
| 29 |
</tbody>
|
| 30 |
</table>
|
| 31 |
|
| 32 |
-
|
| 33 |
|
| 34 |
* コード:https://colab.research.google.com/drive/13m792FEoXszj72viZuBtusYRUL1z6Cu2?usp=sharing
|
| 35 |
* 参考にしたKaggle Notebook : https://www.kaggle.com/code/osanseviero/musiccaps-explorer
|
|
@@ -48,10 +48,24 @@ image = Image.fromarray(np.uint8(D), mode='L') # 'L'は1チャンネルのグ
|
|
| 48 |
image.save('spectrogram_{}.png')
|
| 49 |
```
|
| 50 |
|
| 51 |
-
##
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 52 |
* <font color="red">Subset <b>data 1300-1600</b> and <b>data 3400-3600</b> are not working now, so please get subset_name_list</n>
|
| 53 |
those were removed first</font>.
|
| 54 |
-
### get information about this dataset:
|
| 55 |
```python
|
| 56 |
# Extract dataset's information using huggingface API
|
| 57 |
import requests
|
|
@@ -107,7 +121,7 @@ return subset name list. for example,
|
|
| 107 |
```
|
| 108 |
|
| 109 |
|
| 110 |
-
### load dataset and change to dataloader:
|
| 111 |
* You can use the code below:
|
| 112 |
* <font color="red">...but (;・∀・)I don't know whether this code works efficiently, because I haven't tried this code so far</color>
|
| 113 |
```python
|
|
@@ -181,16 +195,3 @@ def load_datasets():
|
|
| 181 |
train_loader, test_loader = load_datasets()
|
| 182 |
```
|
| 183 |
|
| 184 |
-
### Recover music(wave form) from sprctrogram
|
| 185 |
-
```python
|
| 186 |
-
im = Image.open("pngファイル")
|
| 187 |
-
db_ud = np.uint8(np.array(im))
|
| 188 |
-
amp = librosa.db_to_amplitude(db_ud)
|
| 189 |
-
print(amp.shape)
|
| 190 |
-
# (1025, 861)は20秒のwavファイルをスペクトログラムにした場合
|
| 191 |
-
# (1025, 431)は10秒のwavファイルをスペクトログラムにした場合
|
| 192 |
-
# (1025, 216)は5秒のwavファイルをスペクトログラムにした場合
|
| 193 |
-
|
| 194 |
-
y_inv = librosa.griffinlim(amp*200)
|
| 195 |
-
display(IPython.display.Audio(y_inv, rate=sr))
|
| 196 |
-
```
|
|
|
|
| 9 |
- 10K<n<100K
|
| 10 |
---
|
| 11 |
|
| 12 |
+
# Google/MusicCapsをスペクトログラムにしたデータ。
|
| 13 |
|
| 14 |
+
## Dataset information
|
| 15 |
<table>
|
| 16 |
<thead>
|
| 17 |
<td>画像</td>
|
|
|
|
| 29 |
</tbody>
|
| 30 |
</table>
|
| 31 |
|
| 32 |
+
## How this dataset was made
|
| 33 |
|
| 34 |
* コード:https://colab.research.google.com/drive/13m792FEoXszj72viZuBtusYRUL1z6Cu2?usp=sharing
|
| 35 |
* 参考にしたKaggle Notebook : https://www.kaggle.com/code/osanseviero/musiccaps-explorer
|
|
|
|
| 48 |
image.save('spectrogram_{}.png')
|
| 49 |
```
|
| 50 |
|
| 51 |
+
## Recover music(wave form) from sprctrogram
|
| 52 |
+
```python
|
| 53 |
+
im = Image.open("pngファイル")
|
| 54 |
+
db_ud = np.uint8(np.array(im))
|
| 55 |
+
amp = librosa.db_to_amplitude(db_ud)
|
| 56 |
+
print(amp.shape)
|
| 57 |
+
# (1025, 861)は20秒のwavファイルをスペクトログラムにした場合
|
| 58 |
+
# (1025, 431)は10秒のwavファイルをスペクトログラムにした場合
|
| 59 |
+
# (1025, 216)は5秒のwavファイルをスペクトログラムにした場合
|
| 60 |
+
|
| 61 |
+
y_inv = librosa.griffinlim(amp*200)
|
| 62 |
+
display(IPython.display.Audio(y_inv, rate=sr))
|
| 63 |
+
```
|
| 64 |
+
|
| 65 |
+
## Example : How to use this
|
| 66 |
* <font color="red">Subset <b>data 1300-1600</b> and <b>data 3400-3600</b> are not working now, so please get subset_name_list</n>
|
| 67 |
those were removed first</font>.
|
| 68 |
+
### 1 : get information about this dataset:
|
| 69 |
```python
|
| 70 |
# Extract dataset's information using huggingface API
|
| 71 |
import requests
|
|
|
|
| 121 |
```
|
| 122 |
|
| 123 |
|
| 124 |
+
### 2 : load dataset and change to dataloader:
|
| 125 |
* You can use the code below:
|
| 126 |
* <font color="red">...but (;・∀・)I don't know whether this code works efficiently, because I haven't tried this code so far</color>
|
| 127 |
```python
|
|
|
|
| 195 |
train_loader, test_loader = load_datasets()
|
| 196 |
```
|
| 197 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|