-
Notifications
You must be signed in to change notification settings - Fork 2.1k
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
Merge pull request #5436 from wyh2000/master
Add BibleTTS recipe
- Loading branch information
Showing
23 changed files
with
657 additions
and
0 deletions.
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,113 @@ | ||
# BIBLE-TTS RECIPE | ||
|
||
This is the recipe of single speaker TTS model with [Bible TTS](https://masakhane-io.github.io/bibleTTS/) corpus. | ||
|
||
Our goal is to build a TTS character based system using high-quality BibleTTS dataset. BibleTTS is a large high-quality open Text-to-Speech dataset with up to 80 hours of single speaker. It releases aligned speech and text for six languages spoken in Sub-Saharan Africa. There are two options: | ||
1) Train VITS models from scratch for each language. | ||
|
||
2) Finetune from a pretrained TTS model to accelarate training stage. | ||
|
||
## Recipe flow | ||
|
||
### 1. Data preparation | ||
|
||
Data preparation stage | ||
|
||
Donwload dataset from [Bible TTS](https://masakhane-io.github.io/bibleTTS/), and then run: | ||
|
||
```sh | ||
# Assume that data prep stage (stage 1) is finished | ||
$ ./run.sh --stage 1 --stop-stage 1 | ||
``` | ||
|
||
### 2. VITS training | ||
If you want to train from scratch, run: | ||
```sh | ||
# Specify the language name for training (e.g. Ewe, Hausa, Lingala, Yoruba, Asante-Twi, Akuapem-Twi) | ||
$ lang=Yoruba | ||
$ cd egs2/bibletts/tts1 | ||
$ ./run.sh \ | ||
--lang ${lang} | ||
--train_set ${lang}/tr_no_dev \ | ||
--valid_set ${lang}/dev1 \ | ||
--test_sets "${lang}/dev1 ${lang}/eval1" \ | ||
--srctexts "data/${lang}/tr_no_dev/text" \ | ||
--dumpdir dump/${lang} \ | ||
--stage 2 \ | ||
--ngpu 4 \ | ||
--g2p none \ | ||
--token_type char \ | ||
--cleaner none \ | ||
--tts_task gan_tts \ | ||
--feats_extract linear_spectrogram \ | ||
--feats_normalize none \ | ||
--train_config ./conf/tuning/train_vits.yaml \ | ||
--inference_model train.total_count.ave.pth \ | ||
--tag vits_bibletts_${lang} | ||
|
||
``` | ||
If you want to finetune from a pretrain model, first download a [pretrained model](https://zenodo.org/record/5555690), and then run: | ||
|
||
```sh | ||
# exclude the embedding layer since we are finetuning on a different language | ||
|
||
$ lang=yoruba | ||
$ cd egs2/bibletts/tts1 | ||
$ ./run.sh \ | ||
--train_set ${lang}/tr_no_dev \ | ||
--valid_set ${lang}/dev1 \ | ||
--test_sets "${lang}/dev1 ${lang}/eval1" \ | ||
--srctexts "data/${lang}/tr_no_dev/text" \ | ||
--dumpdir dump/${lang} \ | ||
--stage 2 \ | ||
--ngpu 4 \ | ||
--g2p none \ | ||
--token_type char \ | ||
--cleaner none \ | ||
--tts_task gan_tts \ | ||
--feats_extract linear_spectrogram \ | ||
--feats_normalize none \ | ||
--train_config ./conf/tuning/train_vits.yaml \ | ||
--train_args "--init_param <pretrain-model-path>":::tts.generator.text_encoder.emb \ | ||
--inference_model train.total_count.ave.pth \ | ||
--tag vits_lj_train_phn_1Minter_char_ft_${lang} | ||
|
||
``` | ||
### 3. Inference | ||
|
||
TTS model decoding stage. You can change the decoding setting via --inference_config and --inference_args. | ||
```sh | ||
$ lang=yoruba | ||
$ ./run.sh \ | ||
--train_set ${lang}/tr_no_dev \ | ||
--valid_set ${lang}/dev1 \ | ||
--test_sets "${lang}/dev1 ${lang}/eval1" \ | ||
--srctexts "data/${lang}/tr_no_dev/text" \ | ||
--dumpdir dump/${lang} \ | ||
--ngpu 4 \ | ||
--stage 7 \ | ||
--min_wav_duration 0.38 \ | ||
--g2p none \ | ||
--token_type char \ | ||
--cleaner none \ | ||
--tts_task gan_tts \ | ||
--feats_extract linear_spectrogram \ | ||
--feats_normalize none \ | ||
--train_config conf/tuning/train_vits.yaml \ | ||
--inference_model <checkpoint-**.pth> \ | ||
--tag vits_bibletts_${lang} | ||
``` | ||
|
||
### 4. Objective Evaluation | ||
|
||
```sh | ||
# Evaluate MCD | ||
./pyscripts/utils/evaluate_mcd.py \ | ||
exp/<model_dir_name>/<decode_dir_name>/eval1/wav/wav.scp \ | ||
dump/raw/eval1/wav.scp | ||
|
||
# Evaluate log-F0 RMSE | ||
./pyscripts/utils/evaluate_f0.py \ | ||
exp/<model_dir_name>/<decode_dir_name>/eval1/wav/wav.scp \ | ||
dump/raw/eval1/wav.scp | ||
``` |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,40 @@ | ||
# BIBLE-TTS Results | ||
## Environments | ||
|
||
- python version: `3.8.18` | ||
|
||
- pytorch version: `pytorch 2.0.1+cu117` | ||
|
||
## The details of BibleTTS dataset | ||
|
||
| Languages | Number of Training Utters | Hours (h) | | ||
|:-----: | :-----: | :----: | | ||
| Yoruba | 7,491 | 33.3 | | ||
| Lingala | 11,093 | 71.6 | | ||
| Asante twi | 21,348 | 74.9 | | ||
| Hausa | 40,215 | 86.6 | | ||
| Ewe | 22,192 | 86.8 | | ||
| Akuapem Twi | 24,892 | 67.1 | | ||
|
||
|
||
|
||
|
||
## Results: using objective evaluation (MCD, RMSE) | ||
Pretrained models can be found at [BibleTTS pretrained models](https://huggingface.co/espnet/vits_tts_bibletts_char/tree/main) | ||
|
||
* Experiment results on different languages: | ||
| Languages | MCD | RMSE | | ||
|:-----: | :-----: | :----: | | ||
| Yoruba | 9.66 ± 1.05 | 0.32 ± 0.03 | | ||
| Lingala | 6.98 ± 1.22 | 0.34 ± 0.07 | | ||
| Asante twi | 8.33 ± 1.17 | 0.27 ± 0.04 | | ||
| Hausa | 9.52 ± 1.88 | 0.27 ± 0.07| | ||
| Ewe | 11.11 ± 3.62 | 0.27 ± 0.06 | | ||
| Akuapem Twi | 7.98 ± 1.57 | 0.28 ± 0.06 | | ||
|
||
* Experiment results (Ablation study on Yoruba): | ||
| Settings | MCD_test |MCD_dev | RMSE_test | RMSE_dev | | ||
|:-----: | :-----: | :----: | :----: | :----: | | ||
| Yoruba (train from scratch) | **9.66 ± 1.05** | **6.94 ± 1.96** | **0.32 ± 0.03** | **0.30 ± 0.06** | | ||
| Yoruba (finetune from pretrained model, char-based)) | 10.43 ± 1.13 | 7.97 ± 2.37 | 0.39 ± 0.05 | 0.39 ± 0.10| | ||
| Yoruba (finetune from pretrained model, phn-based) | 10.14 ± 1.26 | 7.42 ± 1.76 | 0.35 ± 0.04 | 0.31 ± 0.07 | |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,110 @@ | ||
# ====== About run.pl, queue.pl, slurm.pl, and ssh.pl ====== | ||
# Usage: <cmd>.pl [options] JOB=1:<nj> <log> <command...> | ||
# e.g. | ||
# run.pl --mem 4G JOB=1:10 echo.JOB.log echo JOB | ||
# | ||
# Options: | ||
# --time <time>: Limit the maximum time to execute. | ||
# --mem <mem>: Limit the maximum memory usage. | ||
# -–max-jobs-run <njob>: Limit the number parallel jobs. This is ignored for non-array jobs. | ||
# --num-threads <ngpu>: Specify the number of CPU core. | ||
# --gpu <ngpu>: Specify the number of GPU devices. | ||
# --config: Change the configuration file from default. | ||
# | ||
# "JOB=1:10" is used for "array jobs" and it can control the number of parallel jobs. | ||
# The left string of "=", i.e. "JOB", is replaced by <N>(Nth job) in the command and the log file name, | ||
# e.g. "echo JOB" is changed to "echo 3" for the 3rd job and "echo 8" for 8th job respectively. | ||
# Note that the number must start with a positive number, so you can't use "JOB=0:10" for example. | ||
# | ||
# run.pl, queue.pl, slurm.pl, and ssh.pl have unified interface, not depending on its backend. | ||
# These options are mapping to specific options for each backend and | ||
# it is configured by "conf/queue.conf" and "conf/slurm.conf" by default. | ||
# If jobs failed, your configuration might be wrong for your environment. | ||
# | ||
# | ||
# The official documentation for run.pl, queue.pl, slurm.pl, and ssh.pl: | ||
# "Parallelization in Kaldi": http://kaldi-asr.org/doc/queue.html | ||
# =========================================================~ | ||
|
||
|
||
# Select the backend used by run.sh from "local", "stdout", "sge", "slurm", or "ssh" | ||
cmd_backend='local' | ||
|
||
# Local machine, without any Job scheduling system | ||
if [ "${cmd_backend}" = local ]; then | ||
|
||
# The other usage | ||
export train_cmd="run.pl" | ||
# Used for "*_train.py": "--gpu" is appended optionally by run.sh | ||
export cuda_cmd="run.pl" | ||
# Used for "*_recog.py" | ||
export decode_cmd="run.pl" | ||
|
||
# Local machine logging to stdout and log file, without any Job scheduling system | ||
elif [ "${cmd_backend}" = stdout ]; then | ||
|
||
# The other usage | ||
export train_cmd="stdout.pl" | ||
# Used for "*_train.py": "--gpu" is appended optionally by run.sh | ||
export cuda_cmd="stdout.pl" | ||
# Used for "*_recog.py" | ||
export decode_cmd="stdout.pl" | ||
|
||
|
||
# "qsub" (Sun Grid Engine, or derivation of it) | ||
elif [ "${cmd_backend}" = sge ]; then | ||
# The default setting is written in conf/queue.conf. | ||
# You must change "-q g.q" for the "queue" for your environment. | ||
# To know the "queue" names, type "qhost -q" | ||
# Note that to use "--gpu *", you have to setup "complex_value" for the system scheduler. | ||
|
||
export train_cmd="queue.pl" | ||
export cuda_cmd="queue.pl" | ||
export decode_cmd="queue.pl" | ||
|
||
|
||
# "qsub" (Torque/PBS.) | ||
elif [ "${cmd_backend}" = pbs ]; then | ||
# The default setting is written in conf/pbs.conf. | ||
|
||
export train_cmd="pbs.pl" | ||
export cuda_cmd="pbs.pl" | ||
export decode_cmd="pbs.pl" | ||
|
||
|
||
# "sbatch" (Slurm) | ||
elif [ "${cmd_backend}" = slurm ]; then | ||
# The default setting is written in conf/slurm.conf. | ||
# You must change "-p cpu" and "-p gpu" for the "partition" for your environment. | ||
# To know the "partion" names, type "sinfo". | ||
# You can use "--gpu * " by default for slurm and it is interpreted as "--gres gpu:*" | ||
# The devices are allocated exclusively using "${CUDA_VISIBLE_DEVICES}". | ||
|
||
export train_cmd="slurm.pl" | ||
export cuda_cmd="slurm.pl" | ||
export decode_cmd="slurm.pl" | ||
|
||
elif [ "${cmd_backend}" = ssh ]; then | ||
# You have to create ".queue/machines" to specify the host to execute jobs. | ||
# e.g. .queue/machines | ||
# host1 | ||
# host2 | ||
# host3 | ||
# Assuming you can login them without any password, i.e. You have to set ssh keys. | ||
|
||
export train_cmd="ssh.pl" | ||
export cuda_cmd="ssh.pl" | ||
export decode_cmd="ssh.pl" | ||
|
||
# This is an example of specifying several unique options in the JHU CLSP cluster setup. | ||
# Users can modify/add their own command options according to their cluster environments. | ||
elif [ "${cmd_backend}" = jhu ]; then | ||
|
||
export train_cmd="queue.pl --mem 2G" | ||
export cuda_cmd="queue-freegpu.pl --mem 2G --gpu 1 --config conf/queue.conf" | ||
export decode_cmd="queue.pl --mem 4G" | ||
|
||
else | ||
echo "$0: Error: Unknown cmd_backend=${cmd_backend}" 1>&2 | ||
return 1 | ||
fi |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,7 @@ | ||
--sample-frequency=16000 | ||
--frame-length=25 # the default is 25 | ||
--low-freq=20 # the default. | ||
--high-freq=7600 # the default is zero meaning use the Nyquist (8k in this case). | ||
--num-mel-bins=30 | ||
--num-ceps=30 | ||
--snip-edges=false |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,11 @@ | ||
# Default configuration | ||
command qsub -V -v PATH -S /bin/bash | ||
option name=* -N $0 | ||
option mem=* -l mem=$0 | ||
option mem=0 # Do not add anything to qsub_opts | ||
option num_threads=* -l ncpus=$0 | ||
option num_threads=1 # Do not add anything to qsub_opts | ||
option num_nodes=* -l nodes=$0:ppn=1 | ||
default gpu=0 | ||
option gpu=0 | ||
option gpu=* -l ngpus=$0 |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,12 @@ | ||
# Default configuration | ||
command qsub -v PATH -cwd -S /bin/bash -j y -l arch=*64* | ||
option name=* -N $0 | ||
option mem=* -l mem_free=$0,ram_free=$0 | ||
option mem=0 # Do not add anything to qsub_opts | ||
option num_threads=* -pe smp $0 | ||
option num_threads=1 # Do not add anything to qsub_opts | ||
option max_jobs_run=* -tc $0 | ||
option num_nodes=* -pe mpi $0 # You must set this PE as allocation_rule=1 | ||
default gpu=0 | ||
option gpu=0 | ||
option gpu=* -l gpu=$0 -q g.q |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,14 @@ | ||
# Default configuration | ||
command sbatch --export=PATH | ||
option name=* --job-name $0 | ||
option time=* --time $0 | ||
option mem=* --mem-per-cpu $0 | ||
option mem=0 | ||
option num_threads=* --cpus-per-task $0 | ||
option num_threads=1 --cpus-per-task 1 | ||
option num_nodes=* --nodes $0 | ||
default gpu=0 | ||
option gpu=0 -p cpu | ||
option gpu=* -p gpu --gres=gpu:$0 -c $0 # Recommend allocating more CPU than, or equal to the number of GPU | ||
# note: the --max-jobs-run option is supported as a special case | ||
# by slurm.pl and you don't have to handle it in the config file. |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,10 @@ | ||
# This configuration is the decoding setting for VITS. | ||
|
||
########################################################## | ||
# DECODING SETTING # | ||
########################################################## | ||
noise_scale: 0.667 # noise scale parameter for the flow in VITS. | ||
noise_scale_dur: 0.8 # noise scale parameter for the stochastic duration predictor in VITS. | ||
speed_control_alpha: 1 # alpha to control the speed of generated speech. | ||
# 1 < alpha makes slower and 1 > alpha makes faster. | ||
use_teacher_forcing: false # whether to use teacher forcing. |
Oops, something went wrong.