Trying to train telugu audio dataset from scratch on vits #3112
Replies: 6 comments 5 replies
-
|
I fixed it by not using phonemes. If anyone wants a detailed explanation, drop a msg. |
Beta Was this translation helpful? Give feedback.
-
|
Can you please provide detailed explanation! |
Beta Was this translation helpful? Give feedback.
-
|
Email me on [email protected] with what you have achieved so far and where you are stuck |
Beta Was this translation helpful? Give feedback.
-
Hi, I a new bee and trying to get my hands dirty in training a new model for telugu if something is not ready. Can you advise how to get telugu in tts? |
Beta Was this translation helpful? Give feedback.
-
|
Could you figure it out?
…On Wed, Jan 29, 2025 at 1:37 PM pschakravarthi ***@***.***> wrote:
I fixed it by not using phonemes. If anyone wants a detailed explanation,
drop a msg.
Hi, I a new bee and trying to get my hands dirty in training a new model
for telugu if something is not ready. Can you advise how to get telugu in
tts?
—
Reply to this email directly, view it on GitHub
<#3112 (reply in thread)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AIAJSQQOGCWFWKHMUMSDZOL2NCD35AVCNFSM6AAAAABN6T6ZJOVHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTCOJZGI3DQMI>
.
You are receiving this because you authored the thread.Message ID:
***@***.***>
|
Beta Was this translation helpful? Give feedback.
-
|
No.. still trying to get something
…On Sat, Apr 5, 2025 at 4:19 PM naveed81 ***@***.***> wrote:
Could you figure it out?
On Wed, Jan 29, 2025 at 1:37 PM pschakravarthi ***@***.***>
wrote:
> I fixed it by not using phonemes. If anyone wants a detailed
explanation,
> drop a msg.
>
> Hi, I a new bee and trying to get my hands dirty in training a new model
> for telugu if something is not ready. Can you advise how to get telugu
in
> tts?
>
> —
> Reply to this email directly, view it on GitHub
> <
#3112 (reply in thread)>,
> or unsubscribe
> <
https://github.com/notifications/unsubscribe-auth/AIAJSQQOGCWFWKHMUMSDZOL2NCD35AVCNFSM6AAAAABN6T6ZJOVHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTCOJZGI3DQMI>
> .
> You are receiving this because you authored the thread.Message ID:
> ***@***.***>
>
—
Reply to this email directly, view it on GitHub
<#3112 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AC5KPFBMSH6T6P4HNFU5VYL2X6YKNAVCNFSM6AAAAABN6T6ZJOVHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTENZTGM4DGNY>
.
You are receiving this because you commented.Message ID:
***@***.***>
|
Beta Was this translation helpful? Give feedback.
-
|
What's your machine and it's configuration
…On Sat, 5 Apr, 2025, 16:37 pschakravarthi, ***@***.***> wrote:
No.. still trying to get something
On Sat, Apr 5, 2025 at 4:19 PM naveed81 ***@***.***> wrote:
> Could you figure it out?
>
> On Wed, Jan 29, 2025 at 1:37 PM pschakravarthi ***@***.***>
> wrote:
>
> > I fixed it by not using phonemes. If anyone wants a detailed
> explanation,
> > drop a msg.
> >
> > Hi, I a new bee and trying to get my hands dirty in training a new
model
> > for telugu if something is not ready. Can you advise how to get telugu
> in
> > tts?
> >
> > —
> > Reply to this email directly, view it on GitHub
> > <
>
#3112 (reply in thread)>,
>
> > or unsubscribe
> > <
>
https://github.com/notifications/unsubscribe-auth/AIAJSQQOGCWFWKHMUMSDZOL2NCD35AVCNFSM6AAAAABN6T6ZJOVHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTCOJZGI3DQMI>
>
> > .
> > You are receiving this because you authored the thread.Message ID:
> > ***@***.***>
> >
>
> —
> Reply to this email directly, view it on GitHub
> <
#3112 (comment)>,
> or unsubscribe
> <
https://github.com/notifications/unsubscribe-auth/AC5KPFBMSH6T6P4HNFU5VYL2X6YKNAVCNFSM6AAAAABN6T6ZJOVHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTENZTGM4DGNY>
> .
> You are receiving this because you commented.Message ID:
> ***@***.***>
>
—
Reply to this email directly, view it on GitHub
<#3112 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AIAJSQXWP5GFFLITV5DOQE32X62NXAVCNFSM6AAAAABN6T6ZJOVHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTENZTGM4TEMY>
.
You are receiving this because you authored the thread.Message ID:
***@***.***>
|
Beta Was this translation helpful? Give feedback.
-
|
I mean, am able to get TTS run in docker with english. But I am trying to see if I can get a model for telugu |
Beta Was this translation helpful? Give feedback.
-
|
I've trained VITS later on.
It worked fine for me
…On Sat, 5 Apr, 2025, 23:41 pschakravarthi, ***@***.***> wrote:
I mean, am able to get TTS run in docker with english. But I am trying to
see if I can get a model for telugu
—
Reply to this email directly, view it on GitHub
<#3112 (reply in thread)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/BBOWQIYLRYBUBBQ5I2AGGYT2YAMDZAVCNFSM6AAAAABN6T6ZJOVHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTENZTGYYDCNY>
.
You are receiving this because you commented.Message ID:
***@***.***>
--
*Engineering **Sciences*
*BVRIT <http://bvrit.ac.in>* | *SVECW*
<http://svecw.edu.in> | *VIT* <http://vishnu.edu.in> | *BVRITH *
<http://bvrithyderabad.edu.in>
*Medical Sciences*
*VDC
<http://vdc.edu.in>* | *SVCP <http://svcp.edu.in>* | *VIPER
<http://viper.ac.in>*| *BVRICE* <http://bvrice.edu.in>*
<http://viper.ac.in>*
|
Beta Was this translation helpful? Give feedback.
-
|
Can you share the model details and how to use it ? |
Beta Was this translation helpful? Give feedback.
-
|
I don't have access to it now
…On Sun, 6 Apr, 2025, 15:45 pschakravarthi, ***@***.***> wrote:
Can you share the model details and how to use it ?
—
Reply to this email directly, view it on GitHub
<#3112 (reply in thread)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AIAJSQV7AE77GLWFGDSM5AL2YD5ETAVCNFSM6AAAAABN6T6ZJOVHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTENZUGAYDENY>
.
You are receiving this because you authored the thread.Message ID:
***@***.***>
|
Beta Was this translation helpful? Give feedback.
Uh oh!
There was an error while loading. Please reload this page.
-
My dataset is a multi speaker one. Below is my training script:
import os
from trainer import Trainer, TrainerArgs
from TTS.tts.configs.shared_configs import BaseDatasetConfig
from TTS.tts.configs.vits_config import VitsConfig
from TTS.tts.datasets import load_tts_samples
from TTS.tts.models.vits import Vits, VitsArgs, VitsAudioConfig
from TTS.tts.utils.speakers import SpeakerManager
from TTS.tts.utils.text.tokenizer import TTSTokenizer
from TTS.utils.audio import AudioProcessor
from TTS.bin.compute_embeddings import compute_embeddings
from TTS.tts.utils.data import get_length_balancer_weights
from TTS.tts.utils.languages import LanguageManager, get_language_balancer_weights
from TTS.tts.utils.speakers import SpeakerManager, get_speaker_balancer_weights, get_speaker_manager
output_path = os.path.dirname(os.path.abspath(file))
dataset_config = BaseDatasetConfig(
formatter="vctk", meta_file_train="", phonemizer="espeak", language="te", path=os.path.join(output_path, "te_male")
)
audio_config = VitsAudioConfig(
sample_rate=22050, win_length=1024, hop_length=256, num_mels=80, mel_fmin=0, mel_fmax=None
)
vitsArgs = VitsArgs(
use_speaker_embedding=True,
)
config = VitsConfig(
model_args=vitsArgs,
audio=audio_config,
run_name="tel_vits",
batch_size=32,
eval_batch_size=16,
batch_group_size=5,
num_loader_workers=0,
num_eval_loader_workers=4,
run_eval=True,
test_delay_epochs=-1,
epochs=1000,
text_cleaner="multilingual_cleaners",
use_phonemes=True,
phoneme_language="te",
# phonemizer="espeak",
phoneme_cache_path=os.path.join(output_path, "phoneme_cache/tel"),
compute_input_seq_cache=True,
print_step=25,
print_eval=False,
mixed_precision=True,
max_text_len=325, # change this if you have a larger VRAM than 16GB
output_path=output_path,
datasets=[dataset_config],
cudnn_benchmark=False,
test_sentences=[
[
"నమస్తే ఖాలిద్ గారు, ఎలా ఉన్నారు?",
"VCTK_tem_00682",
None,
"te",
],
[
"నమస్కారము వెంకటరామణ గారు, వెంకటేశ్వర్లు గారు",
"VCTK_tem_00682",
None,
"te",
],
[
"నవీద్ అహ్మద్ గారు, ఎలా ఉన్నారు?.",
"VCTK_tem_00682",
None,
"te",
],
[
"శుభోదయం ఘట్టమనేని సూర్యప్రకాశ్ గారు",
"VCTK_tem_00682",
None,
"te",
],
],
)
INITIALIZE THE AUDIO PROCESSOR
Audio processor is used for feature extraction and audio I/O.
It mainly serves to the dataloader and the training loggers.
ap = AudioProcessor.init_from_config(config)
INITIALIZE THE TOKENIZER
Tokenizer is used to convert text to sequences of token IDs.
config is updated with the default characters if not defined in the config.
tokenizer, config = TTSTokenizer.init_from_config(config)
LOAD DATA SAMPLES
Each sample is a list of
[text, audio_file_path, speaker_name]You can define your custom sample loader returning the list of samples.
Or define your custom formatter and pass it to the
load_tts_samples.Check
TTS.tts.datasets.load_tts_samplesfor more details.train_samples, eval_samples = load_tts_samples(
dataset_config,
eval_split=True,
eval_split_max_size=config.eval_split_max_size,
eval_split_size=config.eval_split_size,
)
init speaker manager for multi-speaker training
it maps speaker-id to speaker-name in the model and data-loader
speaker_manager = SpeakerManager()
speaker_manager.set_ids_from_data(train_samples + eval_samples, parse_key="speaker_name")
config.model_args.num_speakers = speaker_manager.num_speakers
init model
model = Vits(config, ap, tokenizer, speaker_manager)
init the trainer and 🚀
trainer = Trainer(
TrainerArgs(),
config,
output_path,
model=model,
train_samples=train_samples,
eval_samples=eval_samples,
)
trainer.fit()
So far I have completed 7000 steps and audio I am getting in tensorboard is gibberish (doesnt sound like telugu). Attached samples. What am I doing wrong? Please correct me.
Below are the audio samples after 7000 steps of training. It doesnt sound like Telugu at all, its all gibberish.
https://drive.google.com/drive/folders/1k3OMxE5SpFgV1KQpDpDjMJYKmJhZaAoM?usp=drive_link
This is the config.json file as saved in the run directory:
{
"output_path": "/home/ubuntu/TTS",
"logger_uri": null,
"run_name": "tel_vits",
"project_name": null,
"run_description": "\ud83d\udc38Coqui trainer run.",
"print_step": 25,
"plot_step": 100,
"model_param_stats": false,
"wandb_entity": null,
"dashboard_logger": "tensorboard",
"save_on_interrupt": true,
"log_model_step": 10000,
"save_step": 10000,
"save_n_checkpoints": 5,
"save_checkpoints": true,
"save_all_best": false,
"save_best_after": 10000,
"target_loss": null,
"print_eval": false,
"test_delay_epochs": -1,
"run_eval": true,
"run_eval_steps": null,
"distributed_backend": "nccl",
"distributed_url": "tcp://localhost:54321",
"mixed_precision": true,
"precision": "fp16",
"epochs": 1000,
"batch_size": 32,
"eval_batch_size": 16,
"grad_clip": [
1000,
1000
],
"scheduler_after_epoch": true,
"lr": 0.001,
"optimizer": "AdamW",
"optimizer_params": {
"betas": [
0.8,
0.99
],
"eps": 1e-09,
"weight_decay": 0.01
},
"lr_scheduler": null,
"lr_scheduler_params": {},
"use_grad_scaler": false,
"allow_tf32": false,
"cudnn_enable": true,
"cudnn_deterministic": false,
"cudnn_benchmark": false,
"training_seed": 54321,
"model": "vits",
"num_loader_workers": 0,
"num_eval_loader_workers": 4,
"use_noise_augment": false,
"audio": {
"fft_size": 1024,
"sample_rate": 22050,
"win_length": 1024,
"hop_length": 256,
"num_mels": 80,
"mel_fmin": 0,
"mel_fmax": null
},
"use_phonemes": true,
"phonemizer": "espeak",
"phoneme_language": "te",
"compute_input_seq_cache": true,
"text_cleaner": "multilingual_cleaners",
"enable_eos_bos_chars": false,
"test_sentences_file": "",
"phoneme_cache_path": "/home/ubuntu/TTS/phoneme_cache/tel",
"characters": {
"characters_class": "TTS.tts.utils.text.characters.IPAPhonemes",
"vocab_dict": null,
"pad": "",
"eos": "",
"bos": "",
"blank": "",
"characters": "iy\u0268\u0289\u026fu\u026a\u028f\u028ae\u00f8\u0258\u0259\u0275\u0264o\u025b\u0153\u025c\u025e\u028c\u0254\u00e6\u0250a\u0276\u0251\u0252\u1d7b\u0298\u0253\u01c0\u0257\u01c3\u0284\u01c2\u0260\u01c1\u029bpbtd\u0288\u0256c\u025fk\u0261q\u0262\u0294\u0274\u014b\u0272\u0273n\u0271m\u0299r\u0280\u2c71\u027e\u027d\u0278\u03b2fv\u03b8\u00f0sz\u0283\u0292\u0282\u0290\u00e7\u029dx\u0263\u03c7\u0281\u0127\u0295h\u0266\u026c\u026e\u028b\u0279\u027bj\u0270l\u026d\u028e\u029f\u02c8\u02cc\u02d0\u02d1\u028dw\u0265\u029c\u02a2\u02a1\u0255\u0291\u027a\u0267\u02b2\u025a\u02de\u026b",
"punctuations": "!'(),-.:;? ",
"phonemes": null,
"is_unique": false,
"is_sorted": true
},
"add_blank": true,
"batch_group_size": 5,
"loss_masking": null,
"min_audio_len": 1,
"max_audio_len": Infinity,
"min_text_len": 1,
"max_text_len": 325,
"compute_f0": false,
"compute_energy": false,
"compute_linear_spec": true,
"precompute_num_workers": 0,
"start_by_longest": false,
"shuffle": false,
"drop_last": false,
"datasets": [
{
"formatter": "vctk",
"dataset_name": "",
"path": "/home/ubuntu/TTS/te_male",
"meta_file_train": "",
"ignored_speakers": null,
"language": "te",
"phonemizer": "espeak",
"meta_file_val": "",
"meta_file_attn_mask": ""
}
],
"test_sentences": [
[
"\u0c28\u0c2e\u0c38\u0c4d\u0c24\u0c47 \u0c16\u0c3e\u0c32\u0c3f\u0c26\u0c4d \u0c17\u0c3e\u0c30\u0c41, \u0c0e\u0c32\u0c3e \u0c09\u0c28\u0c4d\u0c28\u0c3e\u0c30\u0c41?",
"VCTK_tem_00682",
null,
"te"
],
[
"\u0c28\u0c2e\u0c38\u0c4d\u0c15\u0c3e\u0c30\u0c2e\u0c41 \u0c35\u0c46\u0c02\u0c15\u0c1f\u0c30\u0c3e\u0c2e\u0c23 \u0c17\u0c3e\u0c30\u0c41, \u0c35\u0c46\u0c02\u0c15\u0c1f\u0c47\u0c36\u0c4d\u0c35\u0c30\u0c4d\u0c32\u0c41 \u0c17\u0c3e\u0c30\u0c41",
"VCTK_tem_00682",
null,
"te"
],
[
"\u0c28\u0c35\u0c40\u0c26\u0c4d \u0c05\u0c39\u0c4d\u0c2e\u0c26\u0c4d \u0c17\u0c3e\u0c30\u0c41, \u0c0e\u0c32\u0c3e \u0c09\u0c28\u0c4d\u0c28\u0c3e\u0c30\u0c41? 2024 \u0c32\u0c4b \u0c15\u0c3e\u0c02\u0c17\u0c4d\u0c30\u0c46\u0c38\u0c4d\u0c15\u0c3f \u0c35\u0c4b\u0c1f\u0c41 \u0c35\u0c47\u0c2f\u0c02\u0c21\u0c3f.",
"VCTK_tem_00682",
null,
"te"
],
[
"\u0c36\u0c41\u0c2d\u0c4b\u0c26\u0c2f\u0c02 \u0c18\u0c1f\u0c4d\u0c1f\u0c2e\u0c28\u0c47\u0c28\u0c3f \u0c38\u0c42\u0c30\u0c4d\u0c2f\u0c2a\u0c4d\u0c30\u0c15\u0c3e\u0c36\u0c4d \u0c17\u0c3e\u0c30\u0c41",
"VCTK_tem_00682",
null,
"te"
]
],
"eval_split_max_size": null,
"eval_split_size": 0.01,
"use_speaker_weighted_sampler": false,
"speaker_weighted_sampler_alpha": 1.0,
"use_language_weighted_sampler": false,
"language_weighted_sampler_alpha": 1.0,
"use_length_weighted_sampler": false,
"length_weighted_sampler_alpha": 1.0,
"model_args": {
"num_chars": 131,
"out_channels": 513,
"spec_segment_size": 32,
"hidden_channels": 192,
"hidden_channels_ffn_text_encoder": 768,
"num_heads_text_encoder": 2,
"num_layers_text_encoder": 6,
"kernel_size_text_encoder": 3,
"dropout_p_text_encoder": 0.1,
"dropout_p_duration_predictor": 0.5,
"kernel_size_posterior_encoder": 5,
"dilation_rate_posterior_encoder": 1,
"num_layers_posterior_encoder": 16,
"kernel_size_flow": 5,
"dilation_rate_flow": 1,
"num_layers_flow": 4,
"resblock_type_decoder": "1",
"resblock_kernel_sizes_decoder": [
3,
7,
11
],
"resblock_dilation_sizes_decoder": [
[
1,
3,
5
],
[
1,
3,
5
],
[
1,
3,
5
]
],
"upsample_rates_decoder": [
8,
8,
2,
2
],
"upsample_initial_channel_decoder": 512,
"upsample_kernel_sizes_decoder": [
16,
16,
4,
4
],
"periods_multi_period_discriminator": [
2,
3,
5,
7,
11
],
"use_sdp": true,
"noise_scale": 1.0,
"inference_noise_scale": 0.667,
"length_scale": 1,
"noise_scale_dp": 1.0,
"inference_noise_scale_dp": 1.0,
"max_inference_len": null,
"init_discriminator": true,
"use_spectral_norm_disriminator": false,
"use_speaker_embedding": true,
"num_speakers": 23,
"speakers_file": "/home/ubuntu/TTS/tel_vits-October-27-2023_03+51AM-99635193/speakers.pth",
"d_vector_file": null,
"speaker_embedding_channels": 256,
"use_d_vector_file": false,
"d_vector_dim": 0,
"detach_dp_input": true,
"use_language_embedding": false,
"embedded_language_dim": 4,
"num_languages": 0,
"language_ids_file": null,
"use_speaker_encoder_as_loss": false,
"speaker_encoder_config_path": "",
"speaker_encoder_model_path": "",
"condition_dp_on_speaker": true,
"freeze_encoder": false,
"freeze_DP": false,
"freeze_PE": false,
"freeze_flow_decoder": false,
"freeze_waveform_decoder": false,
"encoder_sample_rate": null,
"interpolate_z": true,
"reinit_DP": false,
"reinit_text_encoder": false
},
"lr_gen": 0.0002,
"lr_disc": 0.0002,
"lr_scheduler_gen": "ExponentialLR",
"lr_scheduler_gen_params": {
"gamma": 0.999875,
"last_epoch": -1
},
"lr_scheduler_disc": "ExponentialLR",
"lr_scheduler_disc_params": {
"gamma": 0.999875,
"last_epoch": -1
},
"kl_loss_alpha": 1.0,
"disc_loss_alpha": 1.0,
"gen_loss_alpha": 1.0,
"feat_loss_alpha": 1.0,
"mel_loss_alpha": 45.0,
"dur_loss_alpha": 1.0,
"speaker_encoder_loss_alpha": 1.0,
"return_wav": true,
"use_weighted_sampler": false,
"weighted_sampler_attrs": {},
"weighted_sampler_multipliers": {},
"r": 1,
"num_speakers": 0,
"use_speaker_embedding": true,
"speakers_file": "/home/ubuntu/TTS/tel_vits-October-27-2023_03+51AM-99635193/speakers.pth",
"speaker_embedding_channels": 256,
"language_ids_file": null,
"use_language_embedding": false,
"use_d_vector_file": false,
"d_vector_file": null,
"d_vector_dim": 0
}
Beta Was this translation helpful? Give feedback.
All reactions