Datasets:
Update README.md
Browse files
README.md
CHANGED
@@ -47,7 +47,7 @@ configs:
|
|
47 |
[**日本語はこちら**](https://huggingface.co/datasets/joujiboi/japanese-anime-speech-v2/blob/main/README_JA.md)
|
48 |
|
49 |
**japanese-anime-speech-v2** is an audio-text dataset designed for training automatic speech recognition models.
|
50 |
-
The dataset comprises **
|
51 |
This dataset is <b>not</b> an expanded version of <a href="https://huggingface.co/datasets/joujiboi/japanese-anime-speech-v1">japanese-anime-speech-v1</a>.
|
52 |
For that reason, much of the audio from japanese-anime-speech-v1 is **not** included in this dataset.
|
53 |
|
@@ -59,8 +59,7 @@ This genre is characterized by unique vocal and linguistic features and speech p
|
|
59 |
<p>
|
60 |
<b>Content Warning:</b> Please be advised that the audio in this dataset is sourced from visual novels and contains content that is not suitable for all audiences.
|
61 |
<br>
|
62 |
-
|
63 |
-
this dataset has <b>not</b> been filtered for NSFW content. If you use this dataset to train an automatic speech recognition model you *should* clean the dataset first.
|
64 |
</p>
|
65 |
</div>
|
66 |
|
@@ -68,19 +67,22 @@ A list of all audio files and transcriptions can be found [**here**](https://hug
|
|
68 |
|
69 |
# Dataset information
|
70 |
|
71 |
-
* **
|
72 |
-
* **
|
73 |
-
* **
|
|
|
74 |
* Dataset source: **visual novels**
|
75 |
* Audio format: **mp3 (128kbps)**
|
76 |
-
* Latest version: **
|
77 |
-
* Roughly 8% of lines are nsfw
|
78 |
|
79 |
# Changes from V1
|
80 |
|
81 |
* Significantly increased dataset size from 73,004 to 300,506 audio-text pairs
|
82 |
* Changed audio from mp3 (192kbps) -> **mp3 (128kbps)** for better storage efficiency
|
83 |
-
*
|
|
|
|
|
|
|
84 |
|
85 |
# Bias and Limitations
|
86 |
|
@@ -92,8 +94,6 @@ No formatting or cleaning of the transcriptions has been performed (e.g. removin
|
|
92 |
|
93 |
# Future Plans
|
94 |
|
95 |
-
* Create separate splits for safe-for-work and NSFW content
|
96 |
-
* Implement text formatting to improve transcription quality (e.g. removing excessive character repetition)
|
97 |
* Continue expanding the dataset to include more sources
|
98 |
|
99 |
# Use & Credit
|
|
|
47 |
[**日本語はこちら**](https://huggingface.co/datasets/joujiboi/japanese-anime-speech-v2/blob/main/README_JA.md)
|
48 |
|
49 |
**japanese-anime-speech-v2** is an audio-text dataset designed for training automatic speech recognition models.
|
50 |
+
The dataset comprises **292,637 audio clips** and their corresponding transcriptions from various visual novels.
|
51 |
This dataset is <b>not</b> an expanded version of <a href="https://huggingface.co/datasets/joujiboi/japanese-anime-speech-v1">japanese-anime-speech-v1</a>.
|
52 |
For that reason, much of the audio from japanese-anime-speech-v1 is **not** included in this dataset.
|
53 |
|
|
|
59 |
<p>
|
60 |
<b>Content Warning:</b> Please be advised that the audio in this dataset is sourced from visual novels and contains content that is not suitable for all audiences.
|
61 |
<br>
|
62 |
+
An effort was made to split sfw and nsfw content however this was purely based on defined rules that won't be 100% reliable.
|
|
|
63 |
</p>
|
64 |
</div>
|
65 |
|
|
|
67 |
|
68 |
# Dataset information
|
69 |
|
70 |
+
* **292,637** audio-text pairs
|
71 |
+
* **397.54h hours** of sfw audio (86.8%)
|
72 |
+
* **52.36h hours** of nsfw audio (13.2%)
|
73 |
+
* **5.3s** average sfw audio length
|
74 |
* Dataset source: **visual novels**
|
75 |
* Audio format: **mp3 (128kbps)**
|
76 |
+
* Latest version: **V2 - June 29th 2024** (this means the second version of the japanese-anime-speech-v2)
|
|
|
77 |
|
78 |
# Changes from V1
|
79 |
|
80 |
* Significantly increased dataset size from 73,004 to 300,506 audio-text pairs
|
81 |
* Changed audio from mp3 (192kbps) -> **mp3 (128kbps)** for better storage efficiency
|
82 |
+
* sfw and nsfw are on different splits
|
83 |
+
* Repeating characters have been normalised (eg.「はーーーーーーい!」 has become 「はーーい!」)
|
84 |
+
* Removed audio lines containing no dialogue
|
85 |
+
* Removed low quality audio lines
|
86 |
|
87 |
# Bias and Limitations
|
88 |
|
|
|
94 |
|
95 |
# Future Plans
|
96 |
|
|
|
|
|
97 |
* Continue expanding the dataset to include more sources
|
98 |
|
99 |
# Use & Credit
|