Skip to content

Latest commit

 

History

History
112 lines (86 loc) · 5.37 KB

README-SVS-opencpop-pndm.md

File metadata and controls

112 lines (86 loc) · 5.37 KB
Raw

DiffSinger-PNDM

arXiv GitHub Stars downloads

Highlights:

Training diffusion model: 1000 steps

Default pndm_speedup: 40

Inference diffusion model: (1000 / pndm_speedup) steps = 25 steps

You can freely control the inference steps, by adding these arguments in your experiment scripts : --hparams="pndm_speedup=40" or --hparams="pndm_speedup=20" or --hparams="pndm_speedup=10".

Contributed by @luping-liu .

DiffSinger (MIDI SVS | B version | +PNDM)

0. Data Acquirement

For Opencpop dataset: Please strictly follow the instructions of Opencpop. We have no right to give you the access to Opencpop.

The pipeline below is designed for Opencpop dataset:

1. Preparation

Data Preparation

a) Download and extract Opencpop, then create a link to the dataset folder: ln -s /xxx/opencpop data/raw/

b) Run the following scripts to pack the dataset for training/inference.

export PYTHONPATH=.
CUDA_VISIBLE_DEVICES=0 python data_gen/tts/bin/binarize.py --config usr/configs/midi/cascade/opencs/aux_rel.yaml

# `data/binary/opencpop-midi-dp` will be generated.

Vocoder Preparation

We provide the pre-trained model of HifiGAN-Singing which is specially designed for SVS with NSF mechanism.

Also, please unzip pre-trained vocoder and this pendant for vocoder into checkpoints before training your acoustic model.

(Update: You can also move a ckpt with more training steps into this vocoder directory)

This singing vocoder is trained on ~70 hours singing data, which can be viewed as a universal vocoder.

Exp Name Preparation

export MY_DS_EXP_NAME=0831_opencpop_ds1000
.
|--data
    |--raw
        |--opencpop
            |--segments
                |--transcriptions.txt
                |--wavs
|--checkpoints
    |--MY_DS_EXP_NAME (optional)
    |--0109_hifigan_bigpopcs_hop128 (vocoder)
        |--model_ckpt_steps_1512000.ckpt
        |--config.yaml

2. Training Example

CUDA_VISIBLE_DEVICES=0 python tasks/run.py --config usr/configs/midi/e2e/opencpop/ds1000.yaml --exp_name $MY_DS_EXP_NAME --reset

3. Inference from packed test set

CUDA_VISIBLE_DEVICES=0 python tasks/run.py --config usr/configs/midi/e2e/opencpop/ds1000.yaml --exp_name $MY_DS_EXP_NAME --reset --infer

Inference results will be saved in ./checkpoints/MY_DS_EXP_NAME/generated_ by default.

We also provide:

  • the pre-trained model of DiffSinger;

They can be found in here.

Remember to put the pre-trained models in checkpoints directory.

4. Inference from raw inputs

python inference/svs/ds_e2e.py --config usr/configs/midi/e2e/opencpop/ds1000.yaml --exp_name $MY_DS_EXP_NAME

Raw inputs:

inp = {
        'text': '小酒窝长睫毛AP是你最美的记号',
        'notes': 'C#4/Db4 | F#4/Gb4 | G#4/Ab4 | A#4/Bb4 F#4/Gb4 | F#4/Gb4 C#4/Db4 | C#4/Db4 | rest | C#4/Db4 | A#4/Bb4 | G#4/Ab4 | A#4/Bb4 | G#4/Ab4 | F4 | C#4/Db4',
        'notes_duration': '0.407140 | 0.376190 | 0.242180 | 0.509550 0.183420 | 0.315400 0.235020 | 0.361660 | 0.223070 | 0.377270 | 0.340550 | 0.299620 | 0.344510 | 0.283770 | 0.323390 | 0.360340',
        'input_type': 'word'
    }  # user input: Chinese characters
or,
inp = {
        'text': '小酒窝长睫毛AP是你最美的记号',
        'ph_seq': 'x iao j iu w o ch ang ang j ie ie m ao AP sh i n i z ui m ei d e j i h ao',
        'note_seq': 'C#4/Db4 C#4/Db4 F#4/Gb4 F#4/Gb4 G#4/Ab4 G#4/Ab4 A#4/Bb4 A#4/Bb4 F#4/Gb4 F#4/Gb4 F#4/Gb4 C#4/Db4 C#4/Db4 C#4/Db4 rest C#4/Db4 C#4/Db4 A#4/Bb4 A#4/Bb4 G#4/Ab4 G#4/Ab4 A#4/Bb4 A#4/Bb4 G#4/Ab4 G#4/Ab4 F4 F4 C#4/Db4 C#4/Db4',
        'note_dur_seq': '0.407140 0.407140 0.376190 0.376190 0.242180 0.242180 0.509550 0.509550 0.183420 0.315400 0.315400 0.235020 0.361660 0.361660 0.223070 0.377270 0.377270 0.340550 0.340550 0.299620 0.299620 0.344510 0.344510 0.283770 0.283770 0.323390 0.323390 0.360340 0.360340',
        'is_slur_seq': '0 0 0 0 0 0 0 0 1 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0',
        'input_type': 'phoneme'
    }  # input like Opencpop dataset.

Here the inference results will be saved in ./infer_out by default.

5. Some issues.

a) the HifiGAN-Singing is trained on our vocoder dataset and the training set of PopCS. Opencpop is the out-of-domain dataset (unseen speaker). This may cause the deterioration of audio quality, and we are considering fine-tuning this vocoder on the training set of Opencpop.

b) in this version of codes, we used the melody frontend ([lyric + MIDI]->[ph_dur]) to predict phoneme duration. F0 curve is implicitly predicted together with mel-spectrogram.