[TTS] created the finetuning Hifigan 44100Hz recipe on HUI-Audio-Corp…

…us-German. (NVIDIA#4478)

* implemented a script of generating mel-spectrograms for finetuning Hifigan using multiprocessing.
* fixed some typos.

Signed-off-by: Xuesong Yang <1646669+XuesongYang@users.noreply.github.com>
Signed-off-by: Hainan Xu <hainanx@nvidia.com>

examples/tts/conf/hifigan/hifigan_44100.yaml

@@ -18,7 +18,7 @@ highfreq: null
 window: hann
 
 train_n_segments: 16384
-train_max_duration: null
+train_max_duration: null  # change to null to include longer audios.
 train_min_duration: 0.75
 
 val_n_segments: 131072
@@ -70,13 +70,13 @@ model:
 
 trainer:
   num_nodes: 1
-  devices: 1
+  devices: -1
   accelerator: gpu
   strategy: ddp
-  precision: 32
+  precision: 16
   max_steps: ${model.max_steps}
   accumulate_grad_batches: 1
-  enable_checkpointing: False  # Provided by exp_manager
+  enable_checkpointing: false  # Provided by exp_manager
   logger: false # Provided by exp_manager
   log_every_n_steps: 100
   check_val_every_n_epoch: 10

nemo/collections/asr/parts/preprocessing/segment.py

@@ -186,7 +186,7 @@ def segment_from_file(cls, audio_file, target_sr=None, n_segments=0, trim=False,
         try:
             with sf.SoundFile(audio_file, 'r') as f:
                 sample_rate = f.samplerate
-                if n_segments > 0 and len(f) > n_segments:
+                if 0 < n_segments < len(f):
                     max_audio_start = len(f) - n_segments
                     audio_start = random.randint(0, max_audio_start)
                     f.seek(audio_start)

nemo/collections/tts/torch/data.py

@@ -818,15 +818,15 @@ def __init__(
     ):
         """Dataset which can be used for training and fine-tuning vocoder with pre-computed mel-spectrograms.
         Args:
-            manifest_filepath (Union[str, Path, List[str], List[Path]]): Path(s) to the .json manifests containing information on the
-            dataset. Each line in the .json file should be valid json. Note: the .json file itself is not valid
-            json. Each line should contain the following:
+            manifest_filepath (Union[str, Path, List[str], List[Path]]): Path(s) to the .json manifests containing
+            information on the dataset. Each line in the .json file should be valid json. Note: the .json file itself
+            is not valid json. Each line should contain the following:
                 "audio_filepath": <PATH_TO_WAV>,
                 "duration": <Duration of audio clip in seconds> (Optional),
                 "mel_filepath": <PATH_TO_LOG_MEL> (Optional, can be in .npy (numpy.save) or .pt (torch.save) format)
             sample_rate (int): The sample rate of the audio. Or the sample rate that we will resample all files to.
             n_segments (int): The length of audio in samples to load. For example, given a sample rate of 16kHz, and
-                n_segments=16000, a random 1 second section of audio from the clip will be loaded. The section will
+                n_segments=16000, a random 1-second section of audio from the clip will be loaded. The section will
                 be randomly sampled everytime the audio is batched. Can be set to None to load the entire audio.
                 Must be specified if load_precomputed_mel is True.
             max_duration (Optional[float]): Max duration of audio clips in seconds. All samples exceeding this will be
@@ -838,7 +838,8 @@ def __init__(
             ignore_file (Optional[Union[str, Path]]): The location of a pickle-saved list of audio paths
                 that will be pruned prior to training. Defaults to None which does not prune.
             trim (bool): Whether to apply librosa.effects.trim to the audio file. Defaults to False.
-            load_precomputed_mel (bool): Whether to load precomputed mel (useful for fine-tuning). Note: Requires "mel_filepath" to be set in the manifest file.
+            load_precomputed_mel (bool): Whether to load precomputed mel (useful for fine-tuning).
+                Note: Requires "mel_filepath" to be set in the manifest file.
             hop_length (Optional[int]): The hope length between fft computations. Must be specified if load_precomputed_mel is True.
         """
         super().__init__()

nemo/collections/tts/torch/helpers.py

@@ -23,7 +23,7 @@
 class BetaBinomialInterpolator:
     """
         This module calculates alignment prior matrices (based on beta-binomial distribution) using cached popular sizes and image interpolation.
-        The implementation is taken from https://github.com/NVIDIA/DeepLearningExamples.
+        The implementation is taken from https://github.com/NVIDIA/DeepLearningExamples/blob/master/PyTorch/SpeechSynthesis/FastPitch/fastpitch/data_function.py
     """
 
     def __init__(self, round_mel_len_to=50, round_text_len_to=10, cache_size=500):

nemo/core/classes/modelPT.py

@@ -232,7 +232,7 @@ def save_to(self, save_path: str):
 
         .nemo file is an archive (tar.gz) with the following:
             model_config.yaml - model configuration in .yaml format. You can deserialize this into cfg argument for model's constructor
-            model_wights.chpt - model checkpoint
+            model_wights.ckpt - model checkpoint
 
         Args:
             save_path: Path to .nemo file where model instance should be saved
@@ -285,7 +285,7 @@ def restore_from(
                 model as an OmegaConf DictConfig object without instantiating the model.
             trainer: Optional, a pytorch lightning Trainer object that will be forwarded to the
                 instantiated model's constructor.
-            save_restore_connector (SaveRestoreConnector): Can be overrided to add custom save and restore logic.
+            save_restore_connector (SaveRestoreConnector): Can be overridden to add custom save and restore logic.
 
             Example:
                 ```
@@ -331,7 +331,7 @@ def load_from_checkpoint(
     ):
         """
         Loads ModelPT from checkpoint, with some maintenance of restoration.
-        For documentation, please refer to LightningModule.load_from_checkpoin() documentation.
+        For documentation, please refer to LightningModule.load_from_checkpoint() documentation.
         """
         checkpoint = None
         try:

nemo/core/optim/lr_scheduler.py

@@ -868,7 +868,7 @@ def compute_max_steps(
         logging.warning(
             "Please note that drop_last is broken in pytorch 1.6.0. We will fix when pytorch 1.7.0 is released"
         )
-        # TODO: Master verion, not in pytorch 1.6.0
+        # TODO: Master version, not in pytorch 1.6.0
         # sampler_num_samples = math.ceil((num_samples - num_workers)/ num_workers)
 
     steps_per_epoch = _round(sampler_num_samples / batch_size)

scripts/dataset_processing/tts/generate_mels_for_finetune_hifigan.py

@@ -0,0 +1,163 @@
+# Copyright (c) 2022, NVIDIA CORPORATION.  All rights reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+"""
+This script is to generate mel spectrograms from a Fastpitch model checkpoint. Please see general usage below. It runs
+on GPUs by default, but you can add `--num-workers 5 --cpu` as an option to run on CPUs.
+
+$ python scripts/dataset_processing/tts/generate_mels_for_finetune_hifigan.py \
+    --fastpitch-model-ckpt ./models/fastpitch/multi_spk/FastPitch--v_loss\=1.4473-epoch\=209.ckpt \
+    --input-json-manifests /home/xueyang/HUI-Audio-Corpus-German-clean/test_manifest_text_normed_phonemes.json
+    --output-json-manifest-root /home/xueyang/experiments/multi_spk_tts_de
+"""
+
+import argparse
+import json
+from pathlib import Path
+
+import numpy as np
+import soundfile as sf
+import torch
+from joblib import Parallel, delayed
+from tqdm import tqdm
+
+from nemo.collections.tts.models import FastPitchModel
+from nemo.collections.tts.torch.helpers import BetaBinomialInterpolator
+from nemo.utils import logging
+
+
+def get_args():
+    parser = argparse.ArgumentParser(
+        formatter_class=argparse.ArgumentDefaultsHelpFormatter,
+        description="Generate mel spectrograms with pretrained FastPitch model, and create manifests for finetuning Hifigan.",
+    )
+    parser.add_argument(
+        "--fastpitch-model-ckpt", required=True, type=Path, help="Specify a full path of a fastpitch model checkpoint."
+    )
+    parser.add_argument(
+        "--input-json-manifests",
+        nargs="+",
+        required=True,
+        type=Path,
+        help="Specify a full path of a JSON manifest. You could add multiple manifests.",
+    )
+    parser.add_argument(
+        "--output-json-manifest-root",
+        required=True,
+        type=Path,
+        help="Specify a full path of output root that would contain new manifests.",
+    )
+    parser.add_argument(
+        "--num-workers",
+        default=-1,
+        type=int,
+        help="Specify the max number of concurrently Python workers processes. "
+        "If -1 all CPUs are used. If 1 no parallel computing is used.",
+    )
+    parser.add_argument("--cpu", action='store_true', default=False, help="Generate mel spectrograms using CPUs.")
+    args = parser.parse_args()
+    return args
+
+
+def __load_wav(audio_file):
+    with sf.SoundFile(audio_file, 'r') as f:
+        samples = f.read(dtype='float32')
+    return samples.transpose()
+
+
+def __generate_mels(entry, spec_model, device, beta_binomial_interpolator, mel_root):
+    # Generate a spectrograms (we need to use ground truth alignment for correct matching between audio and mels)
+    audio = __load_wav(entry["audio_filepath"])
+    audio = torch.from_numpy(audio).unsqueeze(0).to(device)
+    audio_len = torch.tensor(audio.shape[1], dtype=torch.long, device=device).unsqueeze(0)
+
+    if spec_model.fastpitch.speaker_emb is not None and "speaker" in entry:
+        speaker = torch.tensor([entry['speaker']]).to(device)
+    else:
+        speaker = None
+
+    with torch.no_grad():
+        if "normalized_text" in entry:
+            text = spec_model.parse(entry["normalized_text"], normalize=False)
+        else:
+            text = spec_model.parse(entry['text'])
+
+        text_len = torch.tensor(text.shape[-1], dtype=torch.long, device=device).unsqueeze(0)
+        spect, spect_len = spec_model.preprocessor(input_signal=audio, length=audio_len)
+
+        # Generate attention prior and spectrogram inputs for HiFi-GAN
+        attn_prior = (
+            torch.from_numpy(beta_binomial_interpolator(spect_len.item(), text_len.item()))
+            .unsqueeze(0)
+            .to(text.device)
+        )
+
+        spectrogram = spec_model.forward(
+            text=text, input_lens=text_len, spec=spect, mel_lens=spect_len, attn_prior=attn_prior, speaker=speaker,
+        )[0]
+
+        save_path = mel_root / f"{Path(entry['audio_filepath']).stem}.npy"
+        np.save(save_path, spectrogram[0].to('cpu').numpy())
+        entry["mel_filepath"] = str(save_path)
+
+    return entry
+
+
+def main():
+    args = get_args()
+    ckpt_path = args.fastpitch_model_ckpt
+    input_manifest_filepaths = args.input_json_manifests
+    output_json_manifest_root = args.output_json_manifest_root
+
+    mel_root = output_json_manifest_root / "mels"
+    mel_root.mkdir(exist_ok=True, parents=True)
+
+    # load pretrained FastPitch model checkpoint
+    spec_model = FastPitchModel.load_from_checkpoint(ckpt_path)
+    spec_model.eval()
+    if args.cpu:
+        spec_model.eval()
+    else:
+        spec_model.eval().cuda()
+    device = spec_model.device
+
+    beta_binomial_interpolator = BetaBinomialInterpolator()
+
+    for manifest in input_manifest_filepaths:
+        logging.info(f"Processing {manifest}.")
+        entries = []
+        with open(manifest, "r") as fjson:
+            for line in fjson:
+                entries.append(json.loads(line.strip()))
+
+        if device == "cpu":
+            new_entries = Parallel(n_jobs=args.num_workers)(
+                delayed(__generate_mels)(entry, spec_model, device, beta_binomial_interpolator, mel_root)
+                for entry in entries
+            )
+        else:
+            new_entries = []
+            for entry in tqdm(entries):
+                new_entry = __generate_mels(entry, spec_model, device, beta_binomial_interpolator, mel_root)
+                new_entries.append(new_entry)
+
+        mel_manifest_path = output_json_manifest_root / f"{manifest.stem}_mel{manifest.suffix}"
+        with open(mel_manifest_path, "w") as fmel:
+            for entry in new_entries:
+                fmel.write(json.dumps(entry) + "\n")
+        logging.info(f"Processing {manifest} is complete --> {mel_manifest_path}")
+
+
+if __name__ == "__main__":
+    main()