Skip to content

Commit

Permalink
dataset readme & run.sh
Browse files Browse the repository at this point in the history
  • Loading branch information
@viewfinder-annn
viewfinder-annn committed Dec 29, 2023
1 parent 5b9b7ae commit f0d8003
Show file tree
Hide file tree
Showing 2 changed files with 27 additions and 24 deletions.
34 changes: 18 additions & 16 deletions egs/datasets/README.md
View file
Original file line number Diff line number Diff line change
Expand Up @@ -24,7 +24,9 @@ The downloading link and the file structure tree of each dataset is displayed as

## AudioCaps

AudioCaps is a dataset of around 44K audio-caption pairs, where each audio clip corresponds to a caption with rich semantic information. You can download the dataset [here](https://github.com/cdjkim/audiocaps). The file structure tree is like:
AudioCaps is a dataset of around 44K audio-caption pairs, where each audio clip corresponds to a caption with rich semantic information.

Download AudioCaps dataset [here](https://github.com/cdjkim/audiocaps). The file structure looks like below:

```plaintext
[AudioCaps dataset path]
Expand All @@ -37,7 +39,7 @@ AudioCaps is a dataset of around 44K audio-caption pairs, where each audio clip

## CSD

The official CSD dataset can be download [here](https://zenodo.org/records/4785016). The file structure tree is like:
Download the official CSD dataset [here](https://zenodo.org/records/4785016). The file structure looks like below:

```plaintext
[CSD dataset path]
Expand All @@ -55,7 +57,7 @@ The official CSD dataset can be download [here](https://zenodo.org/records/47850

## CustomSVCDataset

We support custom dataset for Singing Voice Conversion. Arrange your data in the following structure to construct your own dataset:
We support custom dataset for Singing Voice Conversion. Organize your data in the following structure to construct your own dataset:

```plaintext
[Your Custom Dataset Path]
Expand All @@ -72,7 +74,7 @@ We support custom dataset for Singing Voice Conversion. Arrange your data in the

## KiSing

The official KiSing dataset can be download [here](http://shijt.site/index.php/2021/05/16/kising-the-first-open-source-mandarin-singing-voice-synthesis-corpus/). The file structure tree is like:
Download the official KiSing dataset [here](http://shijt.site/index.php/2021/05/16/kising-the-first-open-source-mandarin-singing-voice-synthesis-corpus/). The file structure looks like below:

```plaintext
[KiSing dataset path]
Expand All @@ -84,7 +86,7 @@ The official KiSing dataset can be download [here](http://shijt.site/index.php/2

## LibriTTS

The official LibriTTS dataset can be download [here](https://www.openslr.org/60/). The file structure tree is like:
Download the official LibriTTS dataset [here](https://www.openslr.org/60/). The file structure looks like below:

```plaintext
[LibriTTS dataset path]
Expand Down Expand Up @@ -181,7 +183,7 @@ The official LibriTTS dataset can be download [here](https://www.openslr.org/60/

## LJSpeech

The official LJSpeech dataset can be download [here](https://keithito.com/LJ-Speech-Dataset/). The file structure tree is like:
Download the official LJSpeech dataset [here](https://keithito.com/LJ-Speech-Dataset/). The file structure looks like below:

```plaintext
[LJSpeech dataset path]
Expand All @@ -195,7 +197,7 @@ The official LJSpeech dataset can be download [here](https://keithito.com/LJ-Spe

## M4Singer

The official M4Singer dataset can be downloaded [here](https://drive.google.com/file/d/1xC37E59EWRRFFLdG3aJkVqwtLDgtFNqW/view). The file structure tree is like:
Download the official M4Singer dataset [here](https://drive.google.com/file/d/1xC37E59EWRRFFLdG3aJkVqwtLDgtFNqW/view). The file structure looks like below:

```plaintext
[M4Singer dataset path]
Expand All @@ -214,7 +216,7 @@ The official M4Singer dataset can be downloaded [here](https://drive.google.com/

## NUS-48E

The official NUS-48E dataset can be download [here](https://drive.google.com/drive/folders/12pP9uUl0HTVANU3IPLnumTJiRjPtVUMx). The file structure tree is like:
Download the official NUS-48E dataset [here](https://drive.google.com/drive/folders/12pP9uUl0HTVANU3IPLnumTJiRjPtVUMx). The file structure looks like below:

```plaintext
[NUS-48E dataset path]
Expand All @@ -234,7 +236,7 @@ The official NUS-48E dataset can be download [here](https://drive.google.com/dri

## Opencpop

The official Opera dataset can be downloaded [here](https://wenet.org.cn/opencpop/). The file structure tree is like:
Download the official Opencpop dataset [here](https://wenet.org.cn/opencpop/). The file structure looks like below:

```plaintext
[Opencpop dataset path]
Expand Down Expand Up @@ -268,7 +270,7 @@ The official Opera dataset can be downloaded [here](https://wenet.org.cn/opencpo

## OpenSinger

The official OpenSinger dataset can be downloaded [here](https://drive.google.com/file/d/1EofoZxvalgMjZqzUEuEdleHIZ6SHtNuK/view). The file structure tree is like:
Download the official OpenSinger dataset [here](https://drive.google.com/file/d/1EofoZxvalgMjZqzUEuEdleHIZ6SHtNuK/view). The file structure looks like below:

```plaintext
[OpenSinger dataset path]
Expand All @@ -287,7 +289,7 @@ The official OpenSinger dataset can be downloaded [here](https://drive.google.co

## Opera

The official Opera dataset can be downloaded [here](http://isophonics.net/SingingVoiceDataset). The file structure tree is like:
Download the official Opera dataset [here](http://isophonics.net/SingingVoiceDataset). The file structure looks like below:

```plaintext
[Opera dataset path]
Expand All @@ -306,7 +308,7 @@ The official Opera dataset can be downloaded [here](http://isophonics.net/Singin

## PopBuTFy

The official PopBuTFy dataset can be downloaded [here](https://github.com/MoonInTheRiver/NeuralSVB). The file structure tree is like:
Download the official PopBuTFy dataset [here](https://github.com/MoonInTheRiver/NeuralSVB). The file structure looks like below:

```plaintext
[PopBuTFy dataset path]
Expand All @@ -323,7 +325,7 @@ The official PopBuTFy dataset can be downloaded [here](https://github.com/MoonIn

## PopCS

The official PopCS dataset can be downloaded [here](https://github.com/MoonInTheRiver/DiffSinger/blob/master/resources/apply_form.md). The file structure tree is like:
Download the official PopCS dataset [here](https://github.com/MoonInTheRiver/DiffSinger/blob/master/resources/apply_form.md). The file structure looks like below:

```plaintext
[PopCS dataset path]
Expand All @@ -340,7 +342,7 @@ The official PopCS dataset can be downloaded [here](https://github.com/MoonInThe

## PJS

The official PJS dataset can be downloaded [here](https://sites.google.com/site/shinnosuketakamichi/research-topics/pjs_corpus). The file structure tree is like:
Download the official PJS dataset [here](https://sites.google.com/site/shinnosuketakamichi/research-topics/pjs_corpus). The file structure looks like below:

```plaintext
[PJS dataset path]
Expand All @@ -358,7 +360,7 @@ The official PJS dataset can be downloaded [here](https://sites.google.com/site/

## SVCC

The official SVCC dataset can be downloaded [here](https://github.com/lesterphillip/SVCC23_FastSVC/tree/main/egs/generate_dataset). The file structure tree is like:
Download the official SVCC dataset [here](https://github.com/lesterphillip/SVCC23_FastSVC/tree/main/egs/generate_dataset). The file structure looks like below:

```plaintext
[SVCC dataset path]
Expand All @@ -375,7 +377,7 @@ The official SVCC dataset can be downloaded [here](https://github.com/lesterphil

## VCTK

The official VCTK dataset can be downloaded [here](https://datashare.ed.ac.uk/handle/10283/3443). The file structure tree is like:
Download the official VCTK dataset [here](https://datashare.ed.ac.uk/handle/10283/3443). The file structure looks like below:

```plaintext
[VCTK dataset path]
Expand Down
17 changes: 9 additions & 8 deletions egs/svc/_template/run.sh
View file
Original file line number Diff line number Diff line change
Expand Up @@ -68,14 +68,6 @@ if [ -z "$gpu" ]; then
gpu="0"
fi

if [ -z "$resume_from_ckpt_path" ]; then
resume_from_ckpt_path=""
fi

if [ -z "$resume_type" ]; then
resume_type="resume"
fi

######## Features Extraction ###########
if [ $running_stage -eq 1 ]; then
CUDA_VISIBLE_DEVICES=$gpu python "${work_dir}"/bins/svc/preprocess.py \
Expand All @@ -91,6 +83,15 @@ if [ $running_stage -eq 2 ]; then
fi
echo "Exprimental Name: $exp_name"

# add default value
if [ -z "$resume_from_ckpt_path" ]; then
resume_from_ckpt_path=""
fi

if [ -z "$resume_type" ]; then
resume_type="resume"
fi

if [ "$resume" = true ]; then
echo "Resume from the existing experiment..."
CUDA_VISIBLE_DEVICES="$gpu" accelerate launch "${work_dir}"/bins/svc/train.py \
Expand Down

0 comments on commit f0d8003

Please sign in to comment.