Enforce phonemizer definition for synthesis (#1441)

* Enforce phonemizer definition for synthesis

* Fix train_tts, tokenizer init can now edit config

* Add small change to trigger CI pipeline

* fix wrong output path for one tts_test

* Fix style

* Test config overides by args and tokenizer

* Fix style

TTS/bin/train_tts.py

@@ -57,7 +57,7 @@ def main():
     # init the trainer and 🚀
     trainer = Trainer(
         train_args,
-        config,
+        model.config,
         config.output_path,
         model=model,
         train_samples=train_samples,

TTS/tts/utils/text/tokenizer.py

@@ -191,6 +191,7 @@ class TTSTokenizer:
                     phonemizer = get_phonemizer_by_name(
                         DEF_LANG_TO_PHONEMIZER[config.phoneme_language], **phonemizer_kwargs
                     )
+                    new_config.phonemizer = phonemizer.name()
                 except KeyError as e:
                     raise ValueError(
                         f"""No phonemizer found for language {config.phoneme_language}.

TTS/utils/synthesizer.py

@@ -112,6 +112,9 @@ class Synthesizer(object):
         self.use_phonemes = self.tts_config.use_phonemes
         self.tts_model = setup_tts_model(config=self.tts_config)
 
+        if self.use_phonemes and self.tts_config["phonemizer"] is None:
+            raise ValueError("Phonemizer is not defined in the TTS config.")
+
         if not self.encoder_checkpoint:
             self._set_speaker_encoder_paths_from_tts_config()
 

requirements.txt

@@ -25,7 +25,7 @@ tensorboardX
 pyworld
 # coqui stack
 coqui-trainer
-coqpit                                          # config managemenr
+coqpit # config management
 # chinese g2p deps
 jieba
 pypinyin

tests/tts_tests/test_align_tts_train.py

@@ -1,4 +1,5 @@
 import glob
+import json
 import os
 import shutil
 
@@ -42,7 +43,7 @@ command_train = (
     "--coqpit.datasets.0.meta_file_train metadata.csv "
     "--coqpit.datasets.0.meta_file_val metadata.csv "
     "--coqpit.datasets.0.path tests/data/ljspeech "
-    "--coqpit.test_delay_epochs -1"
+    "--coqpit.test_delay_epochs 0 "
 )
 run_cli(command_train)
 
@@ -54,6 +55,14 @@ continue_config_path = os.path.join(continue_path, "config.json")
 continue_restore_path, _ = get_last_checkpoint(continue_path)
 out_wav_path = os.path.join(get_tests_output_path(), "output.wav")
 
+# Check integrity of the config
+with open(continue_config_path, "r", encoding="utf-8") as f:
+    config_loaded = json.load(f)
+assert config_loaded["characters"] is not None
+assert config_loaded["output_path"] in continue_path
+assert config_loaded["test_delay_epochs"] == 0
+
+# Load the model and run inference
 inference_command = f"CUDA_VISIBLE_DEVICES='{get_device_id()}' tts --text 'This is an example.' --config_path {continue_config_path} --model_path {continue_restore_path} --out_path {out_wav_path}"
 run_cli(inference_command)
 

tests/tts_tests/test_fast_pitch_speaker_emb_train.py

@@ -1,4 +1,5 @@
 import glob
+import json
 import os
 import shutil
 
@@ -74,6 +75,14 @@ out_wav_path = os.path.join(get_tests_output_path(), "output.wav")
 speaker_id = "ljspeech-1"
 continue_speakers_path = os.path.join(continue_path, "speakers.json")
 
+# Check integrity of the config
+with open(continue_config_path, "r", encoding="utf-8") as f:
+    config_loaded = json.load(f)
+assert config_loaded["characters"] is not None
+assert config_loaded["output_path"] in continue_path
+assert config_loaded["test_delay_epochs"] == 0
+
+# Load the model and run inference
 inference_command = f"CUDA_VISIBLE_DEVICES='{get_device_id()}' tts --text 'This is an example.' --speaker_idx {speaker_id} --speakers_file_path {continue_speakers_path} --config_path {continue_config_path} --model_path {continue_restore_path} --out_path {out_wav_path}"
 run_cli(inference_command)
 

tests/tts_tests/test_fast_pitch_train.py

@@ -1,4 +1,5 @@
 import glob
+import json
 import os
 import shutil
 
@@ -73,6 +74,14 @@ continue_config_path = os.path.join(continue_path, "config.json")
 continue_restore_path, _ = get_last_checkpoint(continue_path)
 out_wav_path = os.path.join(get_tests_output_path(), "output.wav")
 
+# Check integrity of the config
+with open(continue_config_path, "r", encoding="utf-8") as f:
+    config_loaded = json.load(f)
+assert config_loaded["characters"] is not None
+assert config_loaded["output_path"] in continue_path
+assert config_loaded["test_delay_epochs"] == 0
+
+# Load the model and run inference
 inference_command = f"CUDA_VISIBLE_DEVICES='{get_device_id()}' tts --text 'This is an example.' --config_path {continue_config_path} --model_path {continue_restore_path} --out_path {out_wav_path}"
 run_cli(inference_command)
 

tests/tts_tests/test_glow_tts_d-vectors_train.py

@@ -1,4 +1,5 @@
 import glob
+import json
 import os
 import shutil
 
@@ -61,6 +62,14 @@ out_wav_path = os.path.join(get_tests_output_path(), "output.wav")
 speaker_id = "ljspeech-1"
 continue_speakers_path = config.d_vector_file
 
+# Check integrity of the config
+with open(continue_config_path, "r", encoding="utf-8") as f:
+    config_loaded = json.load(f)
+assert config_loaded["characters"] is not None
+assert config_loaded["output_path"] in continue_path
+assert config_loaded["test_delay_epochs"] == 0
+
+# Load the model and run inference
 inference_command = f"CUDA_VISIBLE_DEVICES='{get_device_id()}' tts --text 'This is an example.' --speaker_idx {speaker_id} --speakers_file_path {continue_speakers_path} --config_path {continue_config_path} --model_path {continue_restore_path} --out_path {out_wav_path}"
 run_cli(inference_command)
 

tests/tts_tests/test_glow_tts_speaker_emb_train.py

@@ -1,4 +1,5 @@
 import glob
+import json
 import os
 import shutil
 
@@ -58,6 +59,14 @@ out_wav_path = os.path.join(get_tests_output_path(), "output.wav")
 speaker_id = "ljspeech-1"
 continue_speakers_path = os.path.join(continue_path, "speakers.json")
 
+# Check integrity of the config
+with open(continue_config_path, "r", encoding="utf-8") as f:
+    config_loaded = json.load(f)
+assert config_loaded["characters"] is not None
+assert config_loaded["output_path"] in continue_path
+assert config_loaded["test_delay_epochs"] == 0
+
+# Load the model and run inference
 inference_command = f"CUDA_VISIBLE_DEVICES='{get_device_id()}' tts --text 'This is an example.' --speaker_idx {speaker_id} --speakers_file_path {continue_speakers_path} --config_path {continue_config_path} --model_path {continue_restore_path} --out_path {out_wav_path}"
 run_cli(inference_command)
 

tests/tts_tests/test_glow_tts_train.py

@@ -1,4 +1,5 @@
 import glob
+import json
 import os
 import shutil
 
@@ -55,6 +56,14 @@ continue_config_path = os.path.join(continue_path, "config.json")
 continue_restore_path, _ = get_last_checkpoint(continue_path)
 out_wav_path = os.path.join(get_tests_output_path(), "output.wav")
 
+# Check integrity of the config
+with open(continue_config_path, "r", encoding="utf-8") as f:
+    config_loaded = json.load(f)
+assert config_loaded["characters"] is not None
+assert config_loaded["output_path"] in continue_path
+assert config_loaded["test_delay_epochs"] == 0
+
+# Load the model and run inference
 inference_command = f"CUDA_VISIBLE_DEVICES='{get_device_id()}' tts --text 'This is an example.' --config_path {continue_config_path} --model_path {continue_restore_path} --out_path {out_wav_path}"
 run_cli(inference_command)
 

tests/tts_tests/test_speedy_speech_train.py

@@ -1,4 +1,5 @@
 import glob
+import json
 import os
 import shutil
 
@@ -54,6 +55,14 @@ continue_config_path = os.path.join(continue_path, "config.json")
 continue_restore_path, _ = get_last_checkpoint(continue_path)
 out_wav_path = os.path.join(get_tests_output_path(), "output.wav")
 
+# Check integrity of the config
+with open(continue_config_path, "r", encoding="utf-8") as f:
+    config_loaded = json.load(f)
+assert config_loaded["characters"] is not None
+assert config_loaded["output_path"] in continue_path
+assert config_loaded["test_delay_epochs"] == 0
+
+# Load the model and run inference
 inference_command = f"CUDA_VISIBLE_DEVICES='{get_device_id()}' tts --text 'This is an example for it.' --config_path {continue_config_path} --model_path {continue_restore_path} --out_path {out_wav_path}"
 run_cli(inference_command)
 

tests/tts_tests/test_tacotron2_d-vectors_train.py

@@ -1,4 +1,5 @@
 import glob
+import json
 import os
 import shutil
 
@@ -61,6 +62,14 @@ out_wav_path = os.path.join(get_tests_output_path(), "output.wav")
 speaker_id = "ljspeech-1"
 continue_speakers_path = config.d_vector_file
 
+# Check integrity of the config
+with open(continue_config_path, "r", encoding="utf-8") as f:
+    config_loaded = json.load(f)
+assert config_loaded["characters"] is not None
+assert config_loaded["output_path"] in continue_path
+assert config_loaded["test_delay_epochs"] == 0
+
+# Load the model and run inference
 inference_command = f"CUDA_VISIBLE_DEVICES='{get_device_id()}' tts --text 'This is an example.' --speaker_idx {speaker_id} --speakers_file_path {continue_speakers_path} --config_path {continue_config_path} --model_path {continue_restore_path} --out_path {out_wav_path}"
 run_cli(inference_command)
 

tests/tts_tests/test_tacotron2_speaker_emb_train.py

@@ -1,4 +1,5 @@
 import glob
+import json
 import os
 import shutil
 
@@ -59,6 +60,14 @@ out_wav_path = os.path.join(get_tests_output_path(), "output.wav")
 speaker_id = "ljspeech-1"
 continue_speakers_path = os.path.join(continue_path, "speakers.json")
 
+# Check integrity of the config
+with open(continue_config_path, "r", encoding="utf-8") as f:
+    config_loaded = json.load(f)
+assert config_loaded["characters"] is not None
+assert config_loaded["output_path"] in continue_path
+assert config_loaded["test_delay_epochs"] == 0
+
+# Load the model and run inference
 inference_command = f"CUDA_VISIBLE_DEVICES='{get_device_id()}' tts --text 'This is an example.' --speaker_idx {speaker_id} --speakers_file_path {continue_speakers_path} --config_path {continue_config_path} --model_path {continue_restore_path} --out_path {out_wav_path}"
 run_cli(inference_command)
 

tests/tts_tests/test_tacotron2_train.py

@@ -1,4 +1,5 @@
 import glob
+import json
 import os
 import shutil
 
@@ -54,6 +55,14 @@ continue_config_path = os.path.join(continue_path, "config.json")
 continue_restore_path, _ = get_last_checkpoint(continue_path)
 out_wav_path = os.path.join(get_tests_output_path(), "output.wav")
 
+# Check integrity of the config
+with open(continue_config_path, "r", encoding="utf-8") as f:
+    config_loaded = json.load(f)
+assert config_loaded["characters"] is not None
+assert config_loaded["output_path"] in continue_path
+assert config_loaded["test_delay_epochs"] == 0
+
+# Load the model and run inference
 inference_command = f"CUDA_VISIBLE_DEVICES='{get_device_id()}' tts --text 'This is an example.' --config_path {continue_config_path} --model_path {continue_restore_path} --out_path {out_wav_path}"
 run_cli(inference_command)
 

tests/tts_tests/test_tacotron2_train_fsspec_path.py

@@ -1,55 +0,0 @@
-import glob
-import os
-import shutil
-
-from tests import get_device_id, get_tests_output_path, run_cli
-from TTS.tts.configs.tacotron2_config import Tacotron2Config
-
-config_path = os.path.join(get_tests_output_path(), "test_model_config.json")
-output_path = os.path.join(get_tests_output_path(), "train_outputs")
-
-config = Tacotron2Config(
-    r=5,
-    batch_size=8,
-    eval_batch_size=8,
-    num_loader_workers=0,
-    num_eval_loader_workers=0,
-    text_cleaner="english_cleaners",
-    use_phonemes=False,
-    phoneme_language="en-us",
-    phoneme_cache_path=os.path.join(get_tests_output_path(), "train_outputs/phoneme_cache/"),
-    run_eval=True,
-    test_delay_epochs=-1,
-    epochs=1,
-    print_step=1,
-    test_sentences=[
-        "Be a voice, not an echo.",
-    ],
-    print_eval=True,
-    max_decoder_steps=50,
-)
-config.audio.do_trim_silence = True
-config.audio.trim_db = 60
-config.save_json(config_path)
-
-# train the model for one epoch
-command_train = (
-    f"CUDA_VISIBLE_DEVICES='{get_device_id()}' python TTS/bin/train_tts.py --config_path file://{config_path} "
-    f"--coqpit.output_path file://{output_path} "
-    "--coqpit.datasets.0.name ljspeech "
-    "--coqpit.datasets.0.meta_file_train metadata.csv "
-    "--coqpit.datasets.0.meta_file_val metadata.csv "
-    "--coqpit.datasets.0.path tests/data/ljspeech "
-    "--coqpit.test_delay_epochs 0 "
-)
-run_cli(command_train)
-
-# Find latest folder
-continue_path = max(glob.glob(os.path.join(output_path, "*/")), key=os.path.getmtime)
-
-# restore the model and continue training for one more epoch
-command_train = (
-    f"CUDA_VISIBLE_DEVICES='{get_device_id()}' python TTS/bin/train_tts.py --continue_path file://{continue_path} "
-)
-run_cli(command_train)
-shutil.rmtree(continue_path)

tests/tts_tests/test_vits_multilingual_speaker_emb_train.py

@@ -1,4 +1,5 @@
 import glob
+import json
 import os
 import shutil
 
@@ -92,6 +93,14 @@ languae_id = "en"
 continue_speakers_path = os.path.join(continue_path, "speakers.json")
 continue_languages_path = os.path.join(continue_path, "language_ids.json")
 
+# Check integrity of the config
+with open(continue_config_path, "r", encoding="utf-8") as f:
+    config_loaded = json.load(f)
+assert config_loaded["characters"] is not None
+assert config_loaded["output_path"] in continue_path
+assert config_loaded["test_delay_epochs"] == 0
+
+# Load the model and run inference
 inference_command = f"CUDA_VISIBLE_DEVICES='{get_device_id()}' tts --text 'This is an example.' --speaker_idx {speaker_id} --speakers_file_path {continue_speakers_path} --language_ids_file_path {continue_languages_path} --language_idx {languae_id} --config_path {continue_config_path} --model_path {continue_restore_path} --out_path {out_wav_path}"
 run_cli(inference_command)
 

tests/tts_tests/test_vits_multilingual_train-d_vectors.py

@@ -1,4 +1,5 @@
 import glob
+import json
 import os
 import shutil
 
@@ -99,6 +100,14 @@ languae_id = "en"
 continue_speakers_path = config.d_vector_file
 continue_languages_path = os.path.join(continue_path, "language_ids.json")
 
+# Check integrity of the config
+with open(continue_config_path, "r", encoding="utf-8") as f:
+    config_loaded = json.load(f)
+assert config_loaded["characters"] is not None
+assert config_loaded["output_path"] in continue_path
+assert config_loaded["test_delay_epochs"] == 0
+
+# Load the model and run inference
 inference_command = f"CUDA_VISIBLE_DEVICES='{get_device_id()}' tts --text 'This is an example.' --speaker_idx {speaker_id} --speakers_file_path {continue_speakers_path} --language_ids_file_path {continue_languages_path} --language_idx {languae_id} --config_path {continue_config_path} --model_path {continue_restore_path} --out_path {out_wav_path}"
 run_cli(inference_command)
 

tests/tts_tests/test_vits_speaker_emb_train.py

@@ -1,4 +1,5 @@
 import glob
+import json
 import os
 import shutil
 
@@ -65,6 +66,14 @@ out_wav_path = os.path.join(get_tests_output_path(), "output.wav")
 speaker_id = "ljspeech-1"
 continue_speakers_path = os.path.join(continue_path, "speakers.json")
 
+# Check integrity of the config
+with open(continue_config_path, "r", encoding="utf-8") as f:
+    config_loaded = json.load(f)
+assert config_loaded["characters"] is not None
+assert config_loaded["output_path"] in continue_path
+assert config_loaded["test_delay_epochs"] == 0
+
+# Load the model and run inference
 inference_command = f"CUDA_VISIBLE_DEVICES='{get_device_id()}' tts --text 'This is an example.' --speaker_idx {speaker_id} --speakers_file_path {continue_speakers_path} --config_path {continue_config_path} --model_path {continue_restore_path} --out_path {out_wav_path}"
 run_cli(inference_command)
 

tests/tts_tests/test_vits_train.py

@@ -1,4 +1,5 @@
 import glob
+import json
 import os
 import shutil
 
@@ -54,6 +55,14 @@ continue_config_path = os.path.join(continue_path, "config.json")
 continue_restore_path, _ = get_last_checkpoint(continue_path)
 out_wav_path = os.path.join(get_tests_output_path(), "output.wav")
 
+# Check integrity of the config
+with open(continue_config_path, "r", encoding="utf-8") as f:
+    config_loaded = json.load(f)
+assert config_loaded["characters"] is not None
+assert config_loaded["output_path"] in continue_path
+assert config_loaded["test_delay_epochs"] == 0
+
+# Load the model and run inference
 inference_command = f"CUDA_VISIBLE_DEVICES='{get_device_id()}' tts --text 'This is an example.' --config_path {continue_config_path} --model_path {continue_restore_path} --out_path {out_wav_path}"
 run_cli(inference_command)