Codebase for AfroTTS XTTS finetuning

.gitignore

@@ -169,4 +169,5 @@ wandb
 depot/*
 coqui_recipes/*
 local_scripts/*
-coqui_demos/*
+coqui_demos/*
+recipes/ljspeech/xtts_v2/run/*

TTS/tts/datasets/formatters.py

@@ -653,3 +653,15 @@ def bel_tts_formatter(root_path, meta_file, **kwargs):  # pylint: disable=unused
             text = cols[1]
             items.append({"text": text, "audio_file": wav_file, "speaker_name": speaker_name, "root_path": root_path})
     return items
+
+def afrotts(root_path, meta_file, **kwargs):
+    csv_path = os.path.join(root_path, meta_file)
+    csv_file = pd.read_csv(csv_path)
+    csv_file["audio_paths"] = csv_file["audio_paths"].apply(
+        lambda x: x.replace("/AfriSpeech-TTS-D/", root_path)
+    )
+    csv_file = csv_file.rename(columns={"transcript":"text", "audio_paths":"audio_file", "user_ids":"speaker_name"})
+    csv_file['root_path'] = root_path
+    items = csv_file.to_dict('records')
+    return items
+

TTS/tts/layers/xtts/tokenizer.py

@@ -32,7 +32,7 @@ def get_spacy_lang(lang):
         return English()
 
 
-def split_sentence(text, lang, text_split_length=250):
+def split_sentence(text, lang, text_split_length=400):
     """Preprocess the input text"""
     text_splits = []
     if text_split_length is not None and len(text) >= text_split_length:
@@ -595,7 +595,7 @@ def __init__(self, vocab_file=None):
         if vocab_file is not None:
             self.tokenizer = Tokenizer.from_file(vocab_file)
         self.char_limits = {
-            "en": 250,
+            "en": 400,
             "de": 253,
             "fr": 273,
             "es": 239,
@@ -621,7 +621,7 @@ def katsu(self):
 
     def check_input_length(self, txt, lang):
         lang = lang.split("-")[0]  # remove the region
-        limit = self.char_limits.get(lang, 250)
+        limit = self.char_limits.get(lang, 400)
         if len(txt) > limit:
             print(
                 f"[!] Warning: The text length exceeds the character limit of {limit} for language '{lang}', this might cause truncated audio."

recipes/ljspeech/xtts_v2/README.md

@@ -0,0 +1,53 @@
+
+# Coqui AI TTS
+
+This repository contains the code for training a text-to-speech (TTS) model using Coqui AI's TTS framework.
+
+## Installation and Configuration
+
+1. **Clone the repository:**
+
+   ```bash
+   git clone https://github.com/owos/coqui-ai-TTS.git
+   ```
+
+2. **Navigate to the repository root:**
+
+   ```bash
+   cd coqui-ai-TTS
+   ```
+
+3. **Install system dependencies and the code:**
+
+   ```bash
+   make system-deps  # Intended to be used on Ubuntu (Debian). Let us know if you have a different OS.
+   make install
+   ```
+
+4. **Open the following file and redefine the specified variables:**
+
+   File: `recipes/ljspeech/xtts_v2/train_gpt_xtts.py`
+
+   ```python
+   # Line 30
+   path = 'the root path to the audio dirs on your machine'
+
+   # Line 31
+   meta_file_train = "the root path to the train CSV on your machine"
+
+   # Line 32
+   meta_file_val = "the root path to the train CSV on your machine"
+
+   # Line 75
+   SPEAKER_REFERENCE = "a list with a single path to a test audio from the afro tts data"
+   ```
+
+## Running the Code
+
+From the repository root, run the following command:
+
+```bash python
+python3 recipes/ljspeech/xtts_v2/train_gpt_xtts.py
+```
+
+You are now ready to train your TTS model using Coqui AI's framework. Enjoy!

recipes/ljspeech/xtts_v2/train_gpt_xtts.py

@@ -8,7 +8,7 @@
 from TTS.utils.manage import ModelManager
 
 # Logging parameters
-RUN_NAME = "GPT_XTTS_v2.0_LJSpeech_FT"
+RUN_NAME = "GPT_XTTS_v2.0_AfroTTS_FT"
 PROJECT_NAME = "XTTS_trainer"
 DASHBOARD_LOGGER = "tensorboard"
 LOGGER_URI = None
@@ -18,17 +18,18 @@
 
 # Training Parameters
 OPTIMIZER_WD_ONLY_ON_WEIGHTS = True  # for multi-gpu training please make it False
-START_WITH_EVAL = True  # if True it will star with evaluation
-BATCH_SIZE = 3  # set here the batch size
-GRAD_ACUMM_STEPS = 84  # set here the grad accumulation steps
+START_WITH_EVAL = False  # if True it will star with evaluation
+BATCH_SIZE = 16  # set here the batch size
+GRAD_ACUMM_STEPS = 4  # set here the grad accumulation steps
 # Note: we recommend that BATCH_SIZE * GRAD_ACUMM_STEPS need to be at least 252 for more efficient training. You can increase/decrease BATCH_SIZE but then set GRAD_ACUMM_STEPS accordingly.
 
 # Define here the dataset that you want to use for the fine-tuning on.
 config_dataset = BaseDatasetConfig(
-    formatter="ljspeech",
-    dataset_name="ljspeech",
-    path="/raid/datasets/LJSpeech-1.1_24khz/",
-    meta_file_train="/raid/datasets/LJSpeech-1.1_24khz/metadata.csv",
+    formatter="afrotts",
+    dataset_name="afrotts",
+    path="/data4/data/AfriSpeech-TTS-D/",
+    meta_file_train="/data4/abraham/tts/AfriSpeech-TTS/data/afritts-train-clean.csv",
+    meta_file_val="/data4/abraham/tts/AfriSpeech-TTS/data/afritts-dev-clean.csv",
     language="en",
 )
 
@@ -72,7 +73,7 @@
 
 # Training sentences generations
 SPEAKER_REFERENCE = [
-    "./tests/data/ljspeech/wavs/LJ001-0002.wav"  # speaker reference to be used in training test sentences
+    "/data4/data/AfriSpeech-TTS-D/train/1dddeb9f-18ec-4498-b74b-84ac59f2fcf1/e9af9831281555e8685e511f7becdf32_P2L385Vp.wav"  # speaker reference to be used in training test sentences
 ]
 LANGUAGE = config_dataset.language
 
@@ -83,8 +84,8 @@ def main():
         max_conditioning_length=132300,  # 6 secs
         min_conditioning_length=66150,  # 3 secs
         debug_loading_failures=False,
-        max_wav_length=255995,  # ~11.6 seconds
-        max_text_length=200,
+        max_wav_length=255995,  # ~11.6 seconds  661500, #~ 30 seconds #
+        max_text_length=300,
         mel_norm_file=MEL_NORM_FILE,
         dvae_checkpoint=DVAE_CHECKPOINT,
         xtts_checkpoint=XTTS_CHECKPOINT,  # checkpoint path of the model that you want to fine-tune
@@ -110,18 +111,18 @@ def main():
         logger_uri=LOGGER_URI,
         audio=audio_config,
         batch_size=BATCH_SIZE,
-        batch_group_size=48,
+        batch_group_size=64,
         eval_batch_size=BATCH_SIZE,
         num_loader_workers=8,
         eval_split_max_size=256,
         print_step=50,
         plot_step=100,
-        log_model_step=1000,
-        save_step=10000,
-        save_n_checkpoints=1,
+        log_model_step=100,
+        save_step=1000,
+        save_n_checkpoints=3,
         save_checkpoints=True,
         # target_loss="loss",
-        print_eval=False,
+        print_eval=True,
         # Optimizer values like tortoise, pytorch implementation with modifications to not apply WD to non-weight parameters.
         optimizer="AdamW",
         optimizer_wd_only_on_weights=OPTIMIZER_WD_ONLY_ON_WEIGHTS,
@@ -154,7 +155,6 @@ def main():
         eval_split_max_size=config.eval_split_max_size,
         eval_split_size=config.eval_split_size,
     )
-
     # init the trainer and 🚀
     trainer = Trainer(
         TrainerArgs(
@@ -174,3 +174,4 @@ def main():
 
 if __name__ == "__main__":
     main()
+DATASETS_CONFIG_LIST