Merge branch 'reformat' into hifigan-reformat

.circleci/config.yml

@@ -1,51 +0,0 @@
-version: 2
-
-workflows:
-  version: 2
-  test:
-    jobs:
-      - test-3.6
-      - test-3.7
-      - test-3.8
-
-executor: ubuntu-latest
-
-on:
-  push:
-  pull_request:
-    types: [opened, synchronize, reopened]
-
-jobs:
-  test-3.6: &test-template
-    docker:
-      - image: circleci/python:3.6
-    resource_class: large
-    working_directory: ~/repo
-    steps:
-      - checkout
-      - run: |
-          sudo apt update
-          sudo apt install espeak-ng git
-      - run: sudo pip install --upgrade pip
-      - run: sudo pip install -e .
-      - run: |
-          sudo pip install --quiet --upgrade cardboardlint pylint
-          cardboardlinter --refspec ${CIRCLE_BRANCH} -n auto
-      - run: nosetests tests --nocapture --processes=0 --process-timeout=20 --process-restartworker
-      - run: |
-          sudo ./tests/test_glow-tts_train.sh
-          sudo ./tests/test_tacotron_train.sh
-          sudo ./tests/test_vocoder_gan_train.sh
-          sudo ./tests/test_vocoder_wavegrad_train.sh
-          sudo ./tests/test_vocoder_wavernn_train.sh
-          sudo ./tests/test_speedy_speech_train.sh
-
-  test-3.7:
-    <<: *test-template
-    docker:
-      - image: circleci/python:3.7
-
-  test-3.8:
-    <<: *test-template
-    docker:
-      - image: circleci/python:3.8

.github/workflows/main.yml

@@ -35,8 +35,9 @@ jobs:
       - name: Install dependencies
         run: |
           sudo apt update
-          sudo apt install -y espeak-ng git
+          sudo apt install -y git make
           sudo apt install -y python3-wheel gcc
+          make system-deps
       - name: Upgrade pip
         # so we can take advantage of pyproject.toml build-dependency support
         run: python3 -m pip install --upgrade pip
@@ -46,17 +47,6 @@ jobs:
           python3 setup.py egg_info
       - name: Lint check
         run: |
-          cardboardlinter
+          make lint
       - name: Unit tests
-        run: nosetests tests --nocapture --processes=0 --process-timeout=20 --process-restartworker
+        run: make test
-      - name: Test scripts
-        run: |
-          ./tests/test_demo_server.sh
-          ./tests/test_glow-tts_train.sh
-          ./tests/test_tacotron_train.sh
-          ./tests/test_vocoder_gan_train.sh
-          ./tests/test_vocoder_wavegrad_train.sh
-          ./tests/test_vocoder_wavernn_train.sh
-          ./tests/test_speedy_speech_train.sh
-          ./tests/test_resample.sh
-          ./tests/test_compute_statistics.sh

.pre-commit-config.yaml

@@ -0,0 +1,12 @@
+repos:
+  - repo: 'https://github.com/pre-commit/pre-commit-hooks'
+    rev: v2.3.0
+    hooks:
+      - id: check-yaml
+      - id: end-of-file-fixer
+      - id: trailing-whitespace
+  - repo: 'https://github.com/psf/black'
+    rev: 20.8b1
+    hooks:
+      - id: black
+        language_version: python3

CONTRIBUTING.md

@@ -1,51 +1,118 @@
 # Contribution guidelines
 
-This repository is governed by Mozilla's code of conduct and etiquette guidelines. For more details, please read the [Mozilla Community Participation Guidelines](https://www.mozilla.org/about/governance/policies/participation/).
+Welcome to the 🐸TTS!
 
-Before making a Pull Request, check your changes for basic mistakes and style problems by using a linter. We have cardboardlinter setup in this repository, so for example, if you've made some changes and would like to run the linter on just the differences between your work and master, you can use the follow command:
+This repository is governed by the Contributor Covenant Code of Conduct - [CODE_OF_CONDUCT.md](CODE_OF_CONDUCT.md).
 
-```bash
+## Where to start.
-pip install pylint cardboardlint
+We welcome everyone who likes to contribute to 🐸TTS.
-cardboardlinter --refspec master
+You can contribute not only with code but with bug reports, comments, questions, answers, or just a simple tweet to spread the word.
-```
 
-This will compare the code against master and run the linter on all the changes. To run it automatically as a git pre-commit hook, you can do do the following:
+If you like to contribute code, squash a bug but if you don't know where to start, here are some pointers.
 
-```bash
+- [Github Issues Tracker](https://github.com/coqui-ai/TTS/issues)
-cat <<\EOF > .git/hooks/pre-commit
-#!/bin/bash
-if [ ! -x "$(command -v cardboardlinter)" ]; then
-    exit 0
-fi
 
-# First, stash index and work dir, keeping only the
+    This is a place to find feature requests, bugs.
-# to-be-committed changes in the working directory.
-echo "Stashing working tree changes..." 1>&2
-old_stash=$(git rev-parse -q --verify refs/stash)
-git stash save -q --keep-index
-new_stash=$(git rev-parse -q --verify refs/stash)
 
-# If there were no changes (e.g., `--amend` or `--allow-empty`)
+    Issues with the ```good first issue``` tag are good place for beginners to take on.
-# then nothing was stashed, and we should skip everything,
-# including the tests themselves.  (Presumably the tests passed
-# on the previous commit, so there is no need to re-run them.)
-if [ "$old_stash" = "$new_stash" ]; then
-    echo "No changes, skipping lint." 1>&2
-    exit 0
-fi
 
-# Run tests
+- ✨**PR**✨ [pages](https://github.com/coqui-ai/TTS/pulls) with the ```🚀new version``` tag.
-cardboardlinter --refspec HEAD -n auto
-status=$?
 
-# Restore changes
+    We list all the target improvements for the next version. You can pick one of them and start contributing.
-echo "Restoring working tree changes..." 1>&2
-git reset --hard -q && git stash apply --index -q && git stash drop -q
 
-# Exit with status from test-run: nonzero prevents commit
+- Also feel free to suggest new features, ideas and models. We're always open for new things.
-exit $status
+## Sending a ✨**PR**✨
-EOF
-chmod +x .git/hooks/pre-commit
-```
 
-This will run the linters on just the changes made in your commit.
+If you have a new feature, a model to implement, or a bug to squash, go ahead and send a ✨**PR**✨.
+Please use the following steps to send a ✨**PR**✨.
+Let us know if you encounter a problem along the way.
+
+The following steps are tested on an Ubuntu system.
+
+1. Fork 🐸TTS[https://github.com/coqui-ai/TTS] by clicking the fork button at the top right corner of the project page.
+
+2. Clone 🐸TTS and add the main repo as a new remote named ```upsteam```.
+
+    ```bash
+    $ git clone git@github.com:<your Github name>/TTS.git
+    $ cd TTS
+    $ git remote add upstream https://github.com/coqui-ai/TTS.git
+    ```
+
+3. Install 🐸TTS for development.
+
+    ```bash
+    $ make system-deps  # intended to be used on Ubuntu (Debian). Let us know if you have a different OS.
+    $ make install
+    ```
+
+4. Create a new branch with an informative name for your goal.
+
+    ```bash
+    $ git checkout -b an_informative_name_for_my_branch
+    ```
+
+5. Implement your changes on your new branch.
+
+6. Explain your code using [Google Style](https://google.github.io/styleguide/pyguide.html#381-docstrings) docstrings.
+
+7. Add your tests to our test suite under ```tests```  folder. It is important to show that your code works, edge cases are considered, and inform others about the intended use.
+
+8. Run the tests to see how your updates work with the rest of the project. You can repeat this step multiple times as you implement your changes to make sure you are on the right direction.
+
+    ```bash
+    $ make tests
+    ```
+
+9. Format your code. We use ```black``` for code and ```isort``` for ```import``` formatting.
+
+    ```bash
+    $ make style
+    ```
+
+10. Run the linter and correct the issues raised. We use ```pylint``` for linting.  It helps to enforce a coding standard, offers simple refactoring suggestions.
+
+    ```bash
+    $ make lint
+    ```
+
+11. When things are good, add new files and commit your changes.
+
+    ```bash
+    $ git add my_file1.py my_file2.py ...
+    $ git commit
+    ```
+
+    It's a good practice to regularly sync your local copy of the project with the upstream code to keep up with the recent updates.
+
+    ```bash
+    $ git fetch upstream
+    $ git rebase upstream/master
+    # or for the development version
+    $ git rebase upstream/dev
+    ```
+
+12. Send a PR to ```dev``` branch.
+
+    Push your branch to your fork.
+
+    ```bash
+    $ git push -u origin an_informative_name_for_my_branch
+    ```
+
+    Then go to your fork's Github page and click on 'Pull request' to send your ✨**PR**✨.
+
+    Please set ✨**PR**✨'s target branch to ```dev``` as we use ```dev``` to work on the next version.
+
+13. Let's discuss until it is perfect. 💪
+
+    We might ask you for certain changes that would appear in the ✨**PR**✨'s page under 🐸TTS[https://github.com/coqui-ai/TTS/pulls].
+
+14. Once things look perfect, We merge it to the ```dev``` branch and make it ready for the next version.
+
+Feel free to ping us at any step you need help using our communication channels.
+
+If you are new to Github or open-source contribution, These are good resources.
+
+- [Github Docs](https://docs.github.com/en/github/collaborating-with-issues-and-pull-requests/proposing-changes-to-your-work-with-pull-requests)
+- [First-Contribution](https://github.com/firstcontributions/first-contributions)

README.md

@@ -40,6 +40,7 @@ Please use our dedicated channels for questions and discussion. Help is much mor
 | Type                            | Links                               |
 | ------------------------------- | --------------------------------------- |
 | 💾 **Installation**               | [TTS/README.md](https://github.com/coqui-ai/TTS/tree/dev#install-tts)|
+| 👩‍💻 **Contributing**               | [CONTRIBUTING.md](https://github.com/coqui-ai/TTS/blob/main/CONTRIBUTING.md)
 | 👩🏾‍🏫 **Tutorials and Examples**     | [TTS/Wiki](https://github.com/coqui-ai/TTS/wiki/%F0%9F%90%B8-TTS-Notebooks,-Examples-and-Tutorials) |
 | 🚀 **Released Models**            | [TTS Releases](https://github.com/coqui-ai/TTS/releases) and [Experimental Models](https://github.com/coqui-ai/TTS/wiki/Experimental-Released-Models)|
 | 💻 **Docker Image**               | [Repository by @synesthesiam](https://github.com/synesthesiam/docker-coqui-aitts)|
@@ -114,10 +115,19 @@ pip install -e .
 ```
 
 We use ```espeak-ng``` to convert graphemes to phonemes. You might need to install separately.
+
 ```bash
 sudo apt-get install espeak-ng
 ```
 
+If you are on Ubuntu (Debian), you can also run following commands for installation.
+
+```bash
+$ make system-deps  # intended to be used on Ubuntu (Debian). Let us know if you have a diffent OS.
+$ make install
+```
+
+If you are on Windows, 👑@GuyPaddock wrote installation instructions [here](https://stackoverflow.com/questions/66726331/how-can-i-run-mozilla-tts-coqui-tts-training-with-cuda-on-a-windows-system).
 ## Directory Structure
 ```
 |- notebooks/       (Jupyter Notebooks for model evaluation, parameter selection and data analysis.)
@@ -167,11 +177,13 @@ Some of the public datasets that we successfully applied 🐸TTS:
 After the installation, 🐸TTS provides a CLI interface for synthesizing speech using pre-trained models. You can either use your own model or the release models under 🐸TTS.
 
 Listing released 🐸TTS models.
+
 ```bash
 tts --list_models
 ```
 
 Run a tts and a vocoder model from the released model list. (Simply copy and paste the full model names from the list as arguments for the command below.)
+
 ```bash
 tts --text "Text for TTS" \
     --model_name "<type>/<language>/<dataset>/<model_name>" \
@@ -180,6 +192,7 @@ tts --text "Text for TTS" \
 ```
 
 Run your own TTS model (Using Griffin-Lim Vocoder)
+
 ```bash
 tts --text "Text for TTS" \
     --model_path path/to/model.pth.tar \
@@ -188,6 +201,7 @@ tts --text "Text for TTS" \
 ```
 
 Run your own TTS and Vocoder models
+
 ```bash
 tts --text "Text for TTS" \
     --model_path path/to/config.json \
@@ -242,29 +256,7 @@ In case of any error or intercepted execution, if there is no checkpoint yet und
 
 You can also enjoy Tensorboard,  if you point Tensorboard argument```--logdir``` to the experiment folder.
 
-## Contribution guidelines
+## [Contribution guidelines](https://github.com/coqui-ai/TTS/blob/main/CONTRIBUTING.md)
-Please follow the steps below as you send a PR to 🐸. It helps us to keep things organized.
-
-1. Create a new branch.
-2. Implement your changes.
-3. (if applicable) Add [Google Style](https://google.github.io/styleguide/pyguide.html#381-docstrings) docstrings.
-4. (if applicable) Implement a test case under ```tests``` folder.
-5. (Optional but Prefered) Run tests.
-```bash
-./run_tests.sh
-```
-6. Run the ```pylint``` linter.
-```bash
-pip install pylint cardboardlint
-cardboardlinter --refspec master
-```
-7. Send a PR to ```dev``` branch, explain what the change is about.
-8. Let us discuss until we make it perfect :) 💪.
-9. We merge it to the ```dev``` branch once things look good.
-
-Feel free to ping us at any step you need help using our communication channels.
-[Here](https://github.com/firstcontributions/first-contributions) is a good resource for complete beginners.
-
 ### Acknowledgement
 - https://github.com/keithito/tacotron (Dataset pre-processing)
 - https://github.com/r9y9/tacotron_pytorch (Initial Tacotron architecture)

TTS/bin/compute_attention_masks.py

@@ -16,7 +16,7 @@ from TTS.utils.audio import AudioProcessor
 from TTS.utils.io import load_config
 
 if __name__ == "__main__":
-    # pylint: disable=bad-continuation
+    # pylint: disable=bad-option-value
     parser = argparse.ArgumentParser(
         description="""Extract attention masks from trained Tacotron/Tacotron2 models.
 These masks can be used for different purposes including training a TTS model with a Duration Predictor.\n\n"""

TTS/bin/find_unique_chars.py

@@ -7,7 +7,7 @@ from TTS.tts.datasets.preprocess import get_preprocessor_by_name
 
 
 def main():
-    # pylint: disable=bad-continuation
+    # pylint: disable=bad-option-value
     parser = argparse.ArgumentParser(
         description="""Find all the unique characters or phonemes in a dataset.\n\n"""
         """Target dataset must be defined in TTS.tts.datasets.preprocess\n\n"""

TTS/bin/synthesize.py

@@ -23,7 +23,7 @@ def str2bool(v):
 
 
 def main():
-    # pylint: disable=bad-continuation
+    # pylint: disable=bad-option-value
     parser = argparse.ArgumentParser(
         description="""Synthesize speech on command line.\n\n"""
         """You can either use your trained model or choose a model from the provided list.\n\n"""

TTS/bin/train_vocoder_wavernn.py

@@ -24,7 +24,10 @@ from TTS.vocoder.utils.io import save_best_model, save_checkpoint
 
 # from torch.utils.data.distributed import DistributedSampler
 
+<<<<<<< HEAD
 
+=======
+>>>>>>> reformat
 
 use_cuda, num_gpus = setup_torch_training_env(True, True)
 

TTS/speaker_encoder/losses.py

@@ -16,7 +16,7 @@ class GE2ELoss(nn.Module):
             - init_w (float): defines the initial value of w in Equation (5) of [1]
             - init_b (float): definies the initial value of b in Equation (5) of [1]
         """
-        super(GE2ELoss, self).__init__()
+        super().__init__()
         # pylint: disable=E1102
         self.w = nn.Parameter(torch.tensor(init_w))
         # pylint: disable=E1102
@@ -129,7 +129,7 @@ class AngleProtoLoss(nn.Module):
     """
 
     def __init__(self, init_w=10.0, init_b=-5.0):
-        super(AngleProtoLoss, self).__init__()
+        super().__init__()
         # pylint: disable=E1102
         self.w = nn.Parameter(torch.tensor(init_w))
         # pylint: disable=E1102

TTS/tts/datasets/TTSDataset.py

@@ -54,7 +54,7 @@ class MyDataset(Dataset):
             use_noise_augment (bool): enable adding random noise to wav for augmentation.
             verbose (bool): print diagnostic information.
         """
-        super(MyDataset, self).__init__()
+        super().__init__()
         self.batch_group_size = batch_group_size
         self.items = meta_data
         self.outputs_per_step = outputs_per_step

TTS/tts/layers/feed_forward/encoder.py

@@ -136,18 +136,18 @@ class Encoder(nn.Module):
         # init encoder
         if encoder_type.lower() == "relative_position_transformer":
             # text encoder
+            # pylint: disable=unexpected-keyword-arg
             self.encoder = RelativePositionTransformerEncoder(
                 in_hidden_channels, out_channels, in_hidden_channels, encoder_params
-            )  # pylint: disable=unexpected-keyword-arg
+            )
         elif encoder_type.lower() == "residual_conv_bn":
             self.encoder = ResidualConv1dBNEncoder(in_hidden_channels, out_channels, in_hidden_channels, encoder_params)
         elif encoder_type.lower() == "fftransformer":
             assert (
                 in_hidden_channels == out_channels
             ), "[!] must be `in_channels` == `out_channels` when encoder type is 'fftransformer'"
-            self.encoder = FFTransformerBlock(
+            # pylint: disable=unexpected-keyword-arg
-                in_hidden_channels, **encoder_params
+            self.encoder = FFTransformerBlock(in_hidden_channels, **encoder_params)
-            )  # pylint: disable=unexpected-keyword-arg
         else:
             raise NotImplementedError(" [!] unknown encoder type.")
 

TTS/tts/layers/losses.py

@@ -52,7 +52,7 @@ class L1LossMasked(nn.Module):
 
 class MSELossMasked(nn.Module):
     def __init__(self, seq_len_norm):
-        super(MSELossMasked, self).__init__()
+        super().__init__()
         self.seq_len_norm = seq_len_norm
 
     def forward(self, x, target, length):
@@ -132,7 +132,7 @@ class AttentionEntropyLoss(nn.Module):
 
 class BCELossMasked(nn.Module):
     def __init__(self, pos_weight):
-        super(BCELossMasked, self).__init__()
+        super().__init__()
         self.pos_weight = pos_weight
 
     def forward(self, x, target, length):
@@ -193,7 +193,7 @@ class DifferentailSpectralLoss(nn.Module):
 
 class GuidedAttentionLoss(torch.nn.Module):
     def __init__(self, sigma=0.4):
-        super(GuidedAttentionLoss, self).__init__()
+        super().__init__()
         self.sigma = sigma
 
     def _make_ga_masks(self, ilens, olens):
@@ -247,7 +247,7 @@ class TacotronLoss(torch.nn.Module):
     """Collection of Tacotron set-up based on provided config."""
 
     def __init__(self, c, stopnet_pos_weight=10, ga_sigma=0.4):
-        super(TacotronLoss, self).__init__()
+        super().__init__()
         self.stopnet_pos_weight = stopnet_pos_weight
         self.ga_alpha = c.ga_alpha
         self.decoder_diff_spec_alpha = c.decoder_diff_spec_alpha

TTS/tts/layers/tacotron/attentions.py

@@ -16,7 +16,7 @@ class LocationLayer(nn.Module):
     """
 
     def __init__(self, attention_dim, attention_n_filters=32, attention_kernel_size=31):
-        super(LocationLayer, self).__init__()
+        super().__init__()
         self.location_conv1d = nn.Conv1d(
             in_channels=2,
             out_channels=attention_n_filters,
@@ -51,7 +51,7 @@ class GravesAttention(nn.Module):
 
     def __init__(self, query_dim, K):
 
-        super(GravesAttention, self).__init__()
+        super().__init__()
         self._mask_value = 1e-8
         self.K = K
         # self.attention_alignment = 0.05
@@ -178,7 +178,7 @@ class OriginalAttention(nn.Module):
         trans_agent,
         forward_attn_mask,
     ):
-        super(OriginalAttention, self).__init__()
+        super().__init__()
         self.query_layer = Linear(query_dim, attention_dim, bias=False, init_gain="tanh")
         self.inputs_layer = Linear(embedding_dim, attention_dim, bias=False, init_gain="tanh")
         self.v = Linear(attention_dim, 1, bias=True)

TTS/tts/layers/tacotron/common_layers.py

@@ -14,7 +14,7 @@ class Linear(nn.Module):
     """
 
     def __init__(self, in_features, out_features, bias=True, init_gain="linear"):
-        super(Linear, self).__init__()
+        super().__init__()
         self.linear_layer = torch.nn.Linear(in_features, out_features, bias=bias)
         self._init_w(init_gain)
 
@@ -38,7 +38,7 @@ class LinearBN(nn.Module):
     """
 
     def __init__(self, in_features, out_features, bias=True, init_gain="linear"):
-        super(LinearBN, self).__init__()
+        super().__init__()
         self.linear_layer = torch.nn.Linear(in_features, out_features, bias=bias)
         self.batch_normalization = nn.BatchNorm1d(out_features, momentum=0.1, eps=1e-5)
         self._init_w(init_gain)
@@ -87,7 +87,7 @@ class Prenet(nn.Module):
 
     # pylint: disable=dangerous-default-value
     def __init__(self, in_features, prenet_type="original", prenet_dropout=True, out_features=[256, 256], bias=True):
-        super(Prenet, self).__init__()
+        super().__init__()
         self.prenet_type = prenet_type
         self.prenet_dropout = prenet_dropout
         in_features = [in_features] + out_features[:-1]

TTS/tts/layers/tacotron/tacotron.py

@@ -26,7 +26,7 @@ class BatchNormConv1d(nn.Module):
 
     def __init__(self, in_channels, out_channels, kernel_size, stride, padding, activation=None):
 
-        super(BatchNormConv1d, self).__init__()
+        super().__init__()
         self.padding = padding
         self.padder = nn.ConstantPad1d(padding, 0)
         self.conv1d = nn.Conv1d(
@@ -71,7 +71,7 @@ class Highway(nn.Module):
 
     # TODO: Try GLU layer
     def __init__(self, in_features, out_feature):
-        super(Highway, self).__init__()
+        super().__init__()
         self.H = nn.Linear(in_features, out_feature)
         self.H.bias.data.zero_()
         self.T = nn.Linear(in_features, out_feature)
@@ -118,7 +118,7 @@ class CBHG(nn.Module):
         gru_features=128,
         num_highways=4,
     ):
-        super(CBHG, self).__init__()
+        super().__init__()
         self.in_features = in_features
         self.conv_bank_features = conv_bank_features
         self.highway_features = highway_features
@@ -191,7 +191,7 @@ class EncoderCBHG(nn.Module):
     r"""CBHG module with Encoder specific arguments"""
 
     def __init__(self):
-        super(EncoderCBHG, self).__init__()
+        super().__init__()
         self.cbhg = CBHG(
             128,
             K=16,
@@ -217,7 +217,7 @@ class Encoder(nn.Module):
     """
 
     def __init__(self, in_features):
-        super(Encoder, self).__init__()
+        super().__init__()
         self.prenet = Prenet(in_features, out_features=[256, 128])
         self.cbhg = EncoderCBHG()
 
@@ -230,7 +230,7 @@ class Encoder(nn.Module):
 
 class PostCBHG(nn.Module):
     def __init__(self, mel_dim):
-        super(PostCBHG, self).__init__()
+        super().__init__()
         self.cbhg = CBHG(
             mel_dim,
             K=8,
@@ -290,7 +290,7 @@ class Decoder(nn.Module):
         attn_K,
         separate_stopnet,
     ):
-        super(Decoder, self).__init__()
+        super().__init__()
         self.r_init = r
         self.r = r
         self.in_channels = in_channels
@@ -491,7 +491,7 @@ class StopNet(nn.Module):
     """
 
     def __init__(self, in_features):
-        super(StopNet, self).__init__()
+        super().__init__()
         self.dropout = nn.Dropout(0.1)
         self.linear = nn.Linear(in_features, 1)
         torch.nn.init.xavier_uniform_(self.linear.weight, gain=torch.nn.init.calculate_gain("linear"))

TTS/tts/layers/tacotron/tacotron2.py

@@ -24,7 +24,7 @@ class ConvBNBlock(nn.Module):
     """
 
     def __init__(self, in_channels, out_channels, kernel_size, activation=None):
-        super(ConvBNBlock, self).__init__()
+        super().__init__()
         assert (kernel_size - 1) % 2 == 0
         padding = (kernel_size - 1) // 2
         self.convolution1d = nn.Conv1d(in_channels, out_channels, kernel_size, padding=padding)
@@ -57,7 +57,7 @@ class Postnet(nn.Module):
     """
 
     def __init__(self, in_out_channels, num_convs=5):
-        super(Postnet, self).__init__()
+        super().__init__()
         self.convolutions = nn.ModuleList()
         self.convolutions.append(ConvBNBlock(in_out_channels, 512, kernel_size=5, activation="tanh"))
         for _ in range(1, num_convs - 1):
@@ -83,7 +83,7 @@ class Encoder(nn.Module):
     """
 
     def __init__(self, in_out_channels=512):
-        super(Encoder, self).__init__()
+        super().__init__()
         self.convolutions = nn.ModuleList()
         for _ in range(3):
             self.convolutions.append(ConvBNBlock(in_out_channels, in_out_channels, 5, "relu"))
@@ -156,7 +156,7 @@ class Decoder(nn.Module):
         attn_K,
         separate_stopnet,
     ):
-        super(Decoder, self).__init__()
+        super().__init__()
         self.frame_channels = frame_channels
         self.r_init = r
         self.r = r

TTS/tts/models/tacotron.py

@@ -79,7 +79,7 @@ class Tacotron(TacotronAbstract):
         memory_size=5,
         gst_use_speaker_embedding=False,
     ):
-        super(Tacotron, self).__init__(
+        super().__init__(
             num_chars,
             num_speakers,
             r,

TTS/tts/tf/layers/tacotron/common_layers.py

@@ -11,7 +11,7 @@ from tensorflow.python.ops import math_ops
 
 class Linear(keras.layers.Layer):
     def __init__(self, units, use_bias, **kwargs):
-        super(Linear, self).__init__(**kwargs)
+        super().__init__(**kwargs)
         self.linear_layer = keras.layers.Dense(units, use_bias=use_bias, name="linear_layer")
         self.activation = keras.layers.ReLU()
 
@@ -25,7 +25,7 @@ class Linear(keras.layers.Layer):
 
 class LinearBN(keras.layers.Layer):
     def __init__(self, units, use_bias, **kwargs):
-        super(LinearBN, self).__init__(**kwargs)
+        super().__init__(**kwargs)
         self.linear_layer = keras.layers.Dense(units, use_bias=use_bias, name="linear_layer")
         self.batch_normalization = keras.layers.BatchNormalization(
             axis=-1, momentum=0.90, epsilon=1e-5, name="batch_normalization"
@@ -44,7 +44,7 @@ class LinearBN(keras.layers.Layer):
 
 class Prenet(keras.layers.Layer):
     def __init__(self, prenet_type, prenet_dropout, units, bias, **kwargs):
-        super(Prenet, self).__init__(**kwargs)
+        super().__init__(**kwargs)
         self.prenet_type = prenet_type
         self.prenet_dropout = prenet_dropout
         self.linear_layers = []
@@ -98,7 +98,7 @@ class Attention(keras.layers.Layer):
         use_forward_attn_mask,
         **kwargs,
     ):
-        super(Attention, self).__init__(**kwargs)
+        super().__init__(**kwargs)
         self.use_loc_attn = use_loc_attn
         self.loc_attn_n_filters = loc_attn_n_filters
         self.loc_attn_kernel_size = loc_attn_kernel_size
@@ -244,8 +244,7 @@ class Attention(keras.layers.Layer):
 #                  location_attention_filters=32,
 #                  location_attention_kernel_size=31):
 
-#         super(LocationSensitiveAttention,
+#         super(                     self).__init__(units=units,
-#                     self).__init__(units=units,
 #                                     memory=memory,
 #                                     memory_sequence_length=memory_sequence_length,
 #                                     normalize=normalize,

TTS/tts/tf/layers/tacotron/tacotron2.py

@@ -10,7 +10,7 @@ from TTS.tts.tf.utils.tf_utils import shape_list
 # pylint: disable=unexpected-keyword-arg
 class ConvBNBlock(keras.layers.Layer):
     def __init__(self, filters, kernel_size, activation, **kwargs):
-        super(ConvBNBlock, self).__init__(**kwargs)
+        super().__init__(**kwargs)
         self.convolution1d = keras.layers.Conv1D(filters, kernel_size, padding="same", name="convolution1d")
         self.batch_normalization = keras.layers.BatchNormalization(
             axis=2, momentum=0.90, epsilon=1e-5, name="batch_normalization"
@@ -28,7 +28,7 @@ class ConvBNBlock(keras.layers.Layer):
 
 class Postnet(keras.layers.Layer):
     def __init__(self, output_filters, num_convs, **kwargs):
-        super(Postnet, self).__init__(**kwargs)
+        super().__init__(**kwargs)
         self.convolutions = []
         self.convolutions.append(ConvBNBlock(512, 5, "tanh", name="convolutions_0"))
         for idx in range(1, num_convs - 1):
@@ -44,7 +44,7 @@ class Postnet(keras.layers.Layer):
 
 class Encoder(keras.layers.Layer):
     def __init__(self, output_input_dim, **kwargs):
-        super(Encoder, self).__init__(**kwargs)
+        super().__init__(**kwargs)
         self.convolutions = []
         for idx in range(3):
             self.convolutions.append(ConvBNBlock(output_input_dim, 5, "relu", name=f"convolutions_{idx}"))
@@ -81,7 +81,7 @@ class Decoder(keras.layers.Layer):
         enable_tflite,
         **kwargs,
     ):
-        super(Decoder, self).__init__(**kwargs)
+        super().__init__(**kwargs)
         self.frame_dim = frame_dim
         self.r_init = tf.constant(r, dtype=tf.int32)
         self.r = tf.constant(r, dtype=tf.int32)

TTS/tts/tf/models/tacotron2.py

@@ -28,7 +28,11 @@ class Tacotron2(keras.models.Model):
         bidirectional_decoder=False,
         enable_tflite=False,
     ):
+<<<<<<< HEAD
         super(Tacotron2, self).__init__()
+=======
+        super().__init__()
+>>>>>>> reformat
         self.r = r
         self.decoder_output_dim = decoder_output_dim
         self.postnet_output_dim = postnet_output_dim

TTS/tts/utils/speakers.py

@@ -57,7 +57,7 @@ def parse_speakers(c, args, meta_data_train, OUT_PATH):
                 prev_out_path = os.path.dirname(args.restore_path)
                 speaker_mapping = load_speaker_mapping(prev_out_path)
                 speaker_embedding_dim = None
-                assert all([speaker in speaker_mapping for speaker in speakers]), (
+                assert all(speaker in speaker_mapping for speaker in speakers), (
                     "As of now you, you cannot " "introduce new speakers to " "a previously trained model."
                 )
         elif (

TTS/tts/utils/synthesis.py

@@ -1,17 +1,17 @@
 import os
 
-os.environ["TF_CPP_MIN_LOG_LEVEL"] = "3"
+import numpy as np
 import pkg_resources
+import torch
+
+from .text import phoneme_to_sequence, text_to_sequence
+
+os.environ["TF_CPP_MIN_LOG_LEVEL"] = "3"
 
 installed = {pkg.key for pkg in pkg_resources.working_set}  # pylint: disable=not-an-iterable
 if "tensorflow" in installed or "tensorflow-gpu" in installed:
     import tensorflow as tf
 
-import numpy as np
-import torch
-
-from .text import phoneme_to_sequence, text_to_sequence
-
 
 def text_to_seqvec(text, CONFIG):
     text_cleaner = [CONFIG.text_cleaner]

TTS/tts/utils/visual.py

@@ -1,13 +1,13 @@
 import librosa
 import matplotlib
+import matplotlib.pyplot as plt
 import numpy as np
 import torch
 
-matplotlib.use("Agg")
-import matplotlib.pyplot as plt
-
 from TTS.tts.utils.text import phoneme_to_sequence, sequence_to_phoneme
 
+matplotlib.use("Agg")
+
 
 def plot_alignment(alignment, info=None, fig_size=(16, 10), title=None, output_fig=False):
     if isinstance(alignment, torch.Tensor):

TTS/utils/audio.py

@@ -179,8 +179,8 @@ class AudioProcessor(object):
                 S_norm = ((2 * self.max_norm) * S_norm) - self.max_norm
                 if self.clip_norm:
                     S_norm = np.clip(
-                        S_norm, -self.max_norm, self.max_norm
+                        S_norm, -self.max_norm, self.max_norm  # pylint: disable=invalid-unary-operand-type
-                    )  # pylint: disable=invalid-unary-operand-type
+                    )
                 return S_norm
             else:
                 S_norm = self.max_norm * S_norm
@@ -206,8 +206,8 @@ class AudioProcessor(object):
             if self.symmetric_norm:
                 if self.clip_norm:
                     S_denorm = np.clip(
-                        S_denorm, -self.max_norm, self.max_norm
+                        S_denorm, -self.max_norm, self.max_norm  # pylint: disable=invalid-unary-operand-type
-                    )  # pylint: disable=invalid-unary-operand-type
+                    )
                 S_denorm = ((S_denorm + self.max_norm) * -self.min_level_db / (2 * self.max_norm)) + self.min_level_db
                 return S_denorm + self.ref_level_db
             else:
@@ -326,13 +326,7 @@ class AudioProcessor(object):
         )
 
     def _istft(self, y):
-        return librosa.istft(
+        return librosa.istft(y, hop_length=self.hop_length, win_length=self.win_length)
-            y,
-            hop_length=self.hop_length,
-            win_length=self.win_length,
-            window="hann",
-            center=True,
-        )
 
     def _griffin_lim(self, S):
         angles = np.exp(2j * np.pi * np.random.rand(*S.shape))

TTS/utils/distribute.py

@@ -14,7 +14,7 @@ class DistributedSampler(Sampler):
     """
 
     def __init__(self, dataset, num_replicas=None, rank=None):
-        super(DistributedSampler, self).__init__(dataset)
+        super().__init__(dataset)
         if num_replicas is None:
             if not dist.is_available():
                 raise RuntimeError("Requires distributed package to be available")

TTS/utils/io.py

@@ -19,7 +19,7 @@ class AttrDict(dict):
     to class attributes"""
 
     def __init__(self, *args, **kwargs):
-        super(AttrDict, self).__init__(*args, **kwargs)
+        super().__init__(*args, **kwargs)
         self.__dict__ = self
 
 

TTS/utils/radam.py

@@ -25,10 +25,10 @@ class RAdam(Optimizer):
         defaults = dict(
             lr=lr, betas=betas, eps=eps, weight_decay=weight_decay, buffer=[[None, None, None] for _ in range(10)]
         )
-        super(RAdam, self).__init__(params, defaults)
+        super().__init__(params, defaults)
 
     def __setstate__(self, state):  # pylint: disable=useless-super-delegation
-        super(RAdam, self).__setstate__(state)
+        super().__setstate__(state)
 
     def step(self, closure=None):
 

TTS/utils/training.py

@@ -74,7 +74,7 @@ def set_weight_decay(model, weight_decay, skip_list={"decoder.attention.v", "rnn
         if not param.requires_grad:
             continue
 
-        if len(param.shape) == 1 or any([skip_name in name for skip_name in skip_list]):
+        if len(param.shape) == 1 or any((skip_name in name for skip_name in skip_list)):
             no_decay.append(param)
         else:
             decay.append(param)
@@ -85,7 +85,7 @@ def set_weight_decay(model, weight_decay, skip_list={"decoder.attention.v", "rnn
 class NoamLR(torch.optim.lr_scheduler._LRScheduler):
     def __init__(self, optimizer, warmup_steps=0.1, last_epoch=-1):
         self.warmup_steps = float(warmup_steps)
-        super(NoamLR, self).__init__(optimizer, last_epoch)
+        super().__init__(optimizer, last_epoch)
 
     def get_lr(self):
         step = max(self.last_epoch, 1)

TTS/vocoder/datasets/gan_dataset.py

@@ -30,7 +30,7 @@ class GANDataset(Dataset):
         use_cache=False,
         verbose=False,
     ):
-        super(GANDataset, self).__init__()
+        super().__init__()
         self.ap = ap
         self.item_list = items
         self.compute_feat = not isinstance(items[0], (tuple, list))

TTS/vocoder/datasets/wavernn_dataset.py

@@ -22,7 +22,7 @@ class WaveRNNDataset(Dataset):
         verbose=False,
     ):
 
-        super(WaveRNNDataset, self).__init__()
+        super().__init__()
         self.ap = ap
         self.compute_feat = not isinstance(items[0], (tuple, list))
         self.item_list = items

TTS/vocoder/layers/losses.py

@@ -20,8 +20,7 @@ class TorchSTFT(nn.Module):  # pylint: disable=abstract-method
         n_mels=80,
         use_mel=False,
     ):
-        """ Torch based STFT operation """
+        super().__init__()
-        super(TorchSTFT, self).__init__()
         self.n_fft = n_fft
         self.hop_length = hop_length
         self.win_length = win_length
@@ -91,7 +90,7 @@ class STFTLoss(nn.Module):
     It is from ParallelWaveGAN paper https://arxiv.org/pdf/1910.11480.pdf"""
 
     def __init__(self, n_fft, hop_length, win_length):
-        super(STFTLoss, self).__init__()
+        super().__init__()
         self.n_fft = n_fft
         self.hop_length = hop_length
         self.win_length = win_length
@@ -113,7 +112,7 @@ class MultiScaleSTFTLoss(torch.nn.Module):
     It is from ParallelWaveGAN paper https://arxiv.org/pdf/1910.11480.pdf"""
 
     def __init__(self, n_ffts=(1024, 2048, 512), hop_lengths=(120, 240, 50), win_lengths=(600, 1200, 240)):
-        super(MultiScaleSTFTLoss, self).__init__()
+        super().__init__()
         self.loss_funcs = torch.nn.ModuleList()
         for n_fft, hop_length, win_length in zip(n_ffts, hop_lengths, win_lengths):
             self.loss_funcs.append(STFTLoss(n_fft, hop_length, win_length))
@@ -199,7 +198,7 @@ class MSEDLoss(nn.Module):
     def __init__(
         self,
     ):
-        super(MSEDLoss, self).__init__()
+        super().__init__()
         self.loss_func = nn.MSELoss()
 
     # pylint: disable=no-self-use
@@ -225,7 +224,7 @@ class MelganFeatureLoss(nn.Module):
     def __init__(
         self,
     ):
-        super(MelganFeatureLoss, self).__init__()
+        super().__init__()
         self.loss_func = nn.L1Loss()
 
     # pylint: disable=no-self-use

TTS/vocoder/layers/melgan.py

@@ -4,7 +4,7 @@ from torch.nn.utils import weight_norm
 
 class ResidualStack(nn.Module):
     def __init__(self, channels, num_res_blocks, kernel_size):
-        super(ResidualStack, self).__init__()
+        super().__init__()
 
         assert (kernel_size - 1) % 2 == 0, " [!] kernel_size has to be odd."
         base_padding = (kernel_size - 1) // 2

TTS/vocoder/layers/parallel_wavegan.py

@@ -17,7 +17,7 @@ class ResidualBlock(torch.nn.Module):
         bias=True,
         use_causal_conv=False,
     ):
-        super(ResidualBlock, self).__init__()
+        super().__init__()
         self.dropout = dropout
         # no future time stamps available
         if use_causal_conv:

TTS/vocoder/layers/pqmf.py

@@ -8,7 +8,7 @@ from scipy import signal as sig
 # https://github.com/kan-bayashi/ParallelWaveGAN/tree/master/parallel_wavegan
 class PQMF(torch.nn.Module):
     def __init__(self, N=4, taps=62, cutoff=0.15, beta=9.0):
-        super(PQMF, self).__init__()
+        super().__init__()
 
         self.N = N
         self.taps = taps

TTS/vocoder/layers/upsample.py

@@ -4,7 +4,7 @@ from torch.nn import functional as F
 
 class Stretch2d(torch.nn.Module):
     def __init__(self, x_scale, y_scale, mode="nearest"):
-        super(Stretch2d, self).__init__()
+        super().__init__()
         self.x_scale = x_scale
         self.y_scale = y_scale
         self.mode = mode
@@ -28,7 +28,7 @@ class UpsampleNetwork(torch.nn.Module):
         freq_axis_kernel_size=1,
         use_causal_conv=False,
     ):
-        super(UpsampleNetwork, self).__init__()
+        super().__init__()
         self.use_causal_conv = use_causal_conv
         self.up_layers = torch.nn.ModuleList()
         for scale in upsample_factors:
@@ -76,7 +76,7 @@ class ConvUpsample(torch.nn.Module):
         aux_context_window=0,
         use_causal_conv=False,
     ):
-        super(ConvUpsample, self).__init__()
+        super().__init__()
         self.aux_context_window = aux_context_window
         self.use_causal_conv = use_causal_conv and aux_context_window > 0
         # To capture wide-context information in conditional features

TTS/vocoder/models/melgan_discriminator.py

@@ -14,7 +14,7 @@ class MelganDiscriminator(nn.Module):
         downsample_factors=(4, 4, 4, 4),
         groups_denominator=4,
     ):
-        super(MelganDiscriminator, self).__init__()
+        super().__init__()
         self.layers = nn.ModuleList()
 
         layer_kernel_size = np.prod(kernel_sizes)

TTS/vocoder/models/melgan_generator.py

@@ -16,7 +16,7 @@ class MelganGenerator(nn.Module):
         res_kernel=3,
         num_res_blocks=3,
     ):
-        super(MelganGenerator, self).__init__()
+        super().__init__()
 
         # assert model parameters
         assert (proj_kernel - 1) % 2 == 0, " [!] proj_kernel should be an odd number."

TTS/vocoder/models/melgan_multiscale_discriminator.py

@@ -18,7 +18,7 @@ class MelganMultiscaleDiscriminator(nn.Module):
         pooling_padding=2,
         groups_denominator=4,
     ):
-        super(MelganMultiscaleDiscriminator, self).__init__()
+        super().__init__()
 
         self.discriminators = nn.ModuleList(
             [

TTS/vocoder/models/multiband_melgan_generator.py

@@ -15,7 +15,7 @@ class MultibandMelganGenerator(MelganGenerator):
         res_kernel=3,
         num_res_blocks=3,
     ):
-        super(MultibandMelganGenerator, self).__init__(
+        super().__init__(
             in_channels=in_channels,
             out_channels=out_channels,
             proj_kernel=proj_kernel,

TTS/vocoder/models/parallel_wavegan_discriminator.py

@@ -26,7 +26,7 @@ class ParallelWaveganDiscriminator(nn.Module):
         nonlinear_activation_params={"negative_slope": 0.2},
         bias=True,
     ):
-        super(ParallelWaveganDiscriminator, self).__init__()
+        super().__init__()
         assert (kernel_size - 1) % 2 == 0, " [!] does not support even number kernel size."
         assert dilation_factor > 0, " [!] dilation factor must be > 0."
         self.conv_layers = nn.ModuleList()
@@ -100,7 +100,7 @@ class ResidualParallelWaveganDiscriminator(nn.Module):
         nonlinear_activation="LeakyReLU",
         nonlinear_activation_params={"negative_slope": 0.2},
     ):
-        super(ResidualParallelWaveganDiscriminator, self).__init__()
+        super().__init__()
         assert (kernel_size - 1) % 2 == 0, "Not support even number kernel size."
 
         self.in_channels = in_channels

TTS/vocoder/models/parallel_wavegan_generator.py

@@ -33,7 +33,7 @@ class ParallelWaveganGenerator(torch.nn.Module):
         inference_padding=2,
     ):
 
-        super(ParallelWaveganGenerator, self).__init__()
+        super().__init__()
         self.in_channels = in_channels
         self.out_channels = out_channels
         self.aux_channels = aux_channels

TTS/vocoder/models/random_window_discriminator.py

@@ -4,7 +4,7 @@ from torch import nn
 
 class GBlock(nn.Module):
     def __init__(self, in_channels, cond_channels, downsample_factor):
-        super(GBlock, self).__init__()
+        super().__init__()
 
         self.in_channels = in_channels
         self.cond_channels = cond_channels
@@ -35,7 +35,7 @@ class GBlock(nn.Module):
 
 class DBlock(nn.Module):
     def __init__(self, in_channels, out_channels, downsample_factor):
-        super(DBlock, self).__init__()
+        super().__init__()
 
         self.in_channels = in_channels
         self.downsample_factor = downsample_factor
@@ -62,7 +62,7 @@ class DBlock(nn.Module):
 
 class ConditionalDiscriminator(nn.Module):
     def __init__(self, in_channels, cond_channels, downsample_factors=(2, 2, 2), out_channels=(128, 256)):
-        super(ConditionalDiscriminator, self).__init__()
+        super().__init__()
 
         assert len(downsample_factors) == len(out_channels) + 1
 
@@ -106,7 +106,7 @@ class ConditionalDiscriminator(nn.Module):
 
 class UnconditionalDiscriminator(nn.Module):
     def __init__(self, in_channels, base_channels=64, downsample_factors=(8, 4), out_channels=(128, 256)):
-        super(UnconditionalDiscriminator, self).__init__()
+        super().__init__()
 
         self.downsample_factors = downsample_factors
         self.in_channels = in_channels
@@ -148,7 +148,7 @@ class RandomWindowDiscriminator(nn.Module):
         window_sizes=(512, 1024, 2048, 4096, 8192),
     ):
 
-        super(RandomWindowDiscriminator, self).__init__()
+        super().__init__()
         self.cond_channels = cond_channels
         self.window_sizes = window_sizes
         self.hop_length = hop_length

TTS/vocoder/tf/layers/melgan.py

@@ -3,7 +3,7 @@ import tensorflow as tf
 
 class ReflectionPad1d(tf.keras.layers.Layer):
     def __init__(self, padding):
-        super(ReflectionPad1d, self).__init__()
+        super().__init__()
         self.padding = padding
 
     def call(self, x):
@@ -12,7 +12,7 @@ class ReflectionPad1d(tf.keras.layers.Layer):
 
 class ResidualStack(tf.keras.layers.Layer):
     def __init__(self, channels, num_res_blocks, kernel_size, name):
-        super(ResidualStack, self).__init__(name=name)
+        super().__init__(name=name)
 
         assert (kernel_size - 1) % 2 == 0, " [!] kernel_size has to be odd."
         base_padding = (kernel_size - 1) // 2

TTS/vocoder/tf/layers/pqmf.py

@@ -5,7 +5,7 @@ from scipy import signal as sig
 
 class PQMF(tf.keras.layers.Layer):
     def __init__(self, N=4, taps=62, cutoff=0.15, beta=9.0):
-        super(PQMF, self).__init__()
+        super().__init__()
         # define filter coefficient
         self.N = N
         self.taps = taps

TTS/vocoder/tf/models/melgan_generator.py

@@ -1,13 +1,14 @@
 import logging
 import os
 
-os.environ["TF_CPP_MIN_LOG_LEVEL"] = "3"  # FATAL
-logging.getLogger("tensorflow").setLevel(logging.FATAL)
-
 import tensorflow as tf
 
 from TTS.vocoder.tf.layers.melgan import ReflectionPad1d, ResidualStack
 
+os.environ["TF_CPP_MIN_LOG_LEVEL"] = "3"  # FATAL
+logging.getLogger("tensorflow").setLevel(logging.FATAL)
+
+from TTS.vocoder.tf.layers.melgan import ReflectionPad1d, ResidualStack
 
 # pylint: disable=too-many-ancestors
 # pylint: disable=abstract-method
@@ -25,7 +26,7 @@ class MelganGenerator(tf.keras.models.Model):
         res_kernel=3,
         num_res_blocks=3,
     ):
-        super(MelganGenerator, self).__init__()
+        super().__init__()
 
         self.in_channels = in_channels
 

TTS/vocoder/tf/models/multiband_melgan_generator.py

@@ -17,7 +17,7 @@ class MultibandMelganGenerator(MelganGenerator):
         res_kernel=3,
         num_res_blocks=3,
     ):
-        super(MultibandMelganGenerator, self).__init__(
+        super().__init__(
             in_channels=in_channels,
             out_channels=out_channels,
             proj_kernel=proj_kernel,

hubconf.py

@@ -1,8 +1,8 @@
 dependencies = ['torch', 'gdown', 'pysbd', 'phonemizer', 'unidecode', 'pypinyin']  # apt install espeak-ng
 import torch
 
-from TTS.utils.synthesizer import Synthesizer
 from TTS.utils.manage import ModelManager
+from TTS.utils.synthesizer import Synthesizer
 
 
 def tts(model_name='tts_models/en/ljspeech/tacotron2-DCA', vocoder_name=None, use_cuda=False):

requirements.txt

@@ -16,12 +16,14 @@ tqdm
 inflect
 bokeh==1.4.0
 pysbd
-# pyworld
 soundfile
-nose==1.3.7
-cardboardlint==1.3.0
-pylint==2.5.3
 gdown
 umap-learn==0.4.6
 cython
 pyyaml
+# quality and style
+nose
+coverage
+black
+isort
+pylint==2.7.4

requirements_tests.txt

@@ -1,21 +0,0 @@
-torch>=1.5
-tensorflow==2.3.1
-numpy>=1.16.0
-scipy>=0.19.0
-numba==0.48
-librosa==0.7.2
-phonemizer>=2.2.0
-unidecode==0.4.20
-attrdict
-tensorboardX
-matplotlib
-Pillow
-flask
-tqdm
-inflect
-pysbd
-bokeh==1.4.0
-soundfile
-nose==1.3.7
-cardboardlint==1.3.0
-cython

run_bash_tests.sh

@@ -0,0 +1,14 @@
+set -e
+TF_CPP_MIN_LOG_LEVEL=3
+
+# runtime bash based tests
+./tests/bash_tests/test_demo_server.sh && \
+./tests/bash_tests/test_resample.sh && \
+./tests/bash_tests/test_tacotron_train.sh && \
+./tests/bash_tests/test_glow-tts_train.sh && \
+./tests/bash_tests/test_vocoder_gan_train.sh && \
+./tests/bash_tests/test_vocoder_wavernn_train.sh && \
+./tests/bash_tests/test_vocoder_wavegrad_train.sh && \
+./tests/bash_tests/test_speedy_speech_train.sh && \
+./tests/bash_tests/test_aligntts_train.sh && \
+./tests/bash_tests/test_compute_statistics.sh

setup.py

@@ -8,9 +8,8 @@ from distutils.version import LooseVersion
 import numpy
 import setuptools.command.build_py
 import setuptools.command.develop
-from setuptools import setup, Extension, find_packages
 from Cython.Build import cythonize
-
+from setuptools import Extension, find_packages, setup
 
 if LooseVersion(sys.version) < LooseVersion("3.6") or LooseVersion(sys.version) > LooseVersion("3.9"):
     raise RuntimeError(

tests/bash_tests/test_aligntts_train.sh

@@ -3,11 +3,11 @@ set -xe
 BASEDIR=$(dirname "$0")
 echo "$BASEDIR"
 # run training
-CUDA_VISIBLE_DEVICES="" python TTS/bin/train_align_tts.py --config_path $BASEDIR/inputs/test_align_tts.json
+CUDA_VISIBLE_DEVICES="" python TTS/bin/train_align_tts.py --config_path $BASEDIR/../inputs/test_align_tts.json
 # find the training folder
-LATEST_FOLDER=$(ls $BASEDIR/train_outputs/| sort | tail -1)
+LATEST_FOLDER=$(ls $BASEDIR/../train_outputs/| sort | tail -1)
 echo $LATEST_FOLDER
 # continue the previous training
-CUDA_VISIBLE_DEVICES=""  python TTS/bin/train_align_tts.py --continue_path $BASEDIR/train_outputs/$LATEST_FOLDER
+CUDA_VISIBLE_DEVICES=""  python TTS/bin/train_align_tts.py --continue_path $BASEDIR/../train_outputs/$LATEST_FOLDER
 # remove all the outputs
-rm -rf $BASEDIR/train_outputs/
+rm -rf $BASEDIR/../train_outputs/

tests/bash_tests/test_compute_statistics.sh

@@ -3,5 +3,5 @@ set -xe
 BASEDIR=$(dirname "$0")
 echo "$BASEDIR"
 # run training
-CUDA_VISIBLE_DEVICES="" python TTS/bin/compute_statistics.py --config_path $BASEDIR/inputs/test_glow_tts.json --out_path $BASEDIR/outputs/scale_stats.npy
+CUDA_VISIBLE_DEVICES="" python TTS/bin/compute_statistics.py --config_path $BASEDIR/../inputs/test_glow_tts.json --out_path $BASEDIR/../outputs/scale_stats.npy
 

tests/bash_tests/test_glow-tts_train.sh

@@ -3,11 +3,11 @@ set -xe
 BASEDIR=$(dirname "$0")
 echo "$BASEDIR"
 # run training
-CUDA_VISIBLE_DEVICES="" python TTS/bin/train_glow_tts.py --config_path $BASEDIR/inputs/test_glow_tts.json
+CUDA_VISIBLE_DEVICES="" python TTS/bin/train_glow_tts.py --config_path $BASEDIR/../inputs/test_glow_tts.json
 # find the training folder
-LATEST_FOLDER=$(ls $BASEDIR/train_outputs/| sort | tail -1)
+LATEST_FOLDER=$(ls $BASEDIR/../train_outputs/| sort | tail -1)
 echo $LATEST_FOLDER
 # continue the previous training
-CUDA_VISIBLE_DEVICES=""  python TTS/bin/train_glow_tts.py --continue_path $BASEDIR/train_outputs/$LATEST_FOLDER
+CUDA_VISIBLE_DEVICES=""  python TTS/bin/train_glow_tts.py --continue_path $BASEDIR/../train_outputs/$LATEST_FOLDER
 # remove all the outputs
-rm -rf $BASEDIR/train_outputs/
+rm -rf $BASEDIR/../train_outputs/

tests/bash_tests/test_resample.sh

@@ -4,7 +4,7 @@ BASEDIR=$(dirname "$0")
 TARGET_SR=16000
 echo "$BASEDIR"
 #run the resample script
-python TTS/bin/resample.py --input_dir $BASEDIR/data/ljspeech --output_dir $BASEDIR/outputs/resample_tests --output_sr $TARGET_SR
+python TTS/bin/resample.py --input_dir $BASEDIR/../data/ljspeech --output_dir $BASEDIR/outputs/resample_tests --output_sr $TARGET_SR
 #check samplerate of output
 OUT_SR=$( (echo "import librosa" ; echo "y, sr = librosa.load('"$BASEDIR"/outputs/resample_tests/wavs/LJ001-0012.wav', sr=None)" ; echo "print(sr)") | python )
 OUT_SR=$(($OUT_SR + 0))

tests/bash_tests/test_speedy_speech_train.sh

@@ -3,11 +3,11 @@ set -xe
 BASEDIR=$(dirname "$0")
 echo "$BASEDIR"
 # run training
-CUDA_VISIBLE_DEVICES="" python TTS/bin/train_speedy_speech.py --config_path $BASEDIR/inputs/test_speedy_speech.json
+CUDA_VISIBLE_DEVICES="" python TTS/bin/train_speedy_speech.py --config_path $BASEDIR/../inputs/test_speedy_speech.json
 # find the training folder
-LATEST_FOLDER=$(ls $BASEDIR/train_outputs/| sort | tail -1)
+LATEST_FOLDER=$(ls $BASEDIR/../train_outputs/| sort | tail -1)
 echo $LATEST_FOLDER
 # continue the previous training
-CUDA_VISIBLE_DEVICES=""  python TTS/bin/train_speedy_speech.py --continue_path $BASEDIR/train_outputs/$LATEST_FOLDER
+CUDA_VISIBLE_DEVICES=""  python TTS/bin/train_speedy_speech.py --continue_path $BASEDIR/../train_outputs/$LATEST_FOLDER
 # remove all the outputs
-rm -rf $BASEDIR/train_outputs/
+rm -rf $BASEDIR/../train_outputs/

tests/bash_tests/test_tacotron_train.sh

@@ -4,33 +4,33 @@ BASEDIR=$(dirname "$0")
 echo "$BASEDIR"
 
 # run training
-CUDA_VISIBLE_DEVICES="" python TTS/bin/train_tacotron.py --config_path $BASEDIR/inputs/test_tacotron_config.json
+CUDA_VISIBLE_DEVICES="" python TTS/bin/train_tacotron.py --config_path $BASEDIR/../inputs/test_tacotron_config.json
 # find the training folder
-LATEST_FOLDER=$(ls $BASEDIR/train_outputs/| sort | tail -1)
+LATEST_FOLDER=$(ls $BASEDIR/../train_outputs/| sort | tail -1)
 echo $LATEST_FOLDER
 # continue the previous training
-CUDA_VISIBLE_DEVICES=""  python TTS/bin/train_tacotron.py --continue_path $BASEDIR/train_outputs/$LATEST_FOLDER
+CUDA_VISIBLE_DEVICES=""  python TTS/bin/train_tacotron.py --continue_path $BASEDIR/../train_outputs/$LATEST_FOLDER
 # remove all the outputs
-rm -rf $BASEDIR/train_outputs/
+rm -rf $BASEDIR/../train_outputs/
 
 # run Tacotron bi-directional decoder
-CUDA_VISIBLE_DEVICES="" python TTS/bin/train_tacotron.py --config_path $BASEDIR/inputs/test_tacotron_bd_config.json
+CUDA_VISIBLE_DEVICES="" python TTS/bin/train_tacotron.py --config_path $BASEDIR/../inputs/test_tacotron_bd_config.json
 # find the training folder
-LATEST_FOLDER=$(ls $BASEDIR/train_outputs/| sort | tail -1)
+LATEST_FOLDER=$(ls $BASEDIR/../train_outputs/| sort | tail -1)
 echo $LATEST_FOLDER
 # continue the previous training
-CUDA_VISIBLE_DEVICES=""  python TTS/bin/train_tacotron.py --continue_path $BASEDIR/train_outputs/$LATEST_FOLDER
+CUDA_VISIBLE_DEVICES=""  python TTS/bin/train_tacotron.py --continue_path $BASEDIR/../train_outputs/$LATEST_FOLDER
 # remove all the outputs
-rm -rf $BASEDIR/train_outputs/
+rm -rf $BASEDIR/../train_outputs/
 
 # Tacotron2
 # run training
-CUDA_VISIBLE_DEVICES="" python TTS/bin/train_tacotron.py --config_path $BASEDIR/inputs/test_tacotron2_config.json
+CUDA_VISIBLE_DEVICES="" python TTS/bin/train_tacotron.py --config_path $BASEDIR/../inputs/test_tacotron2_config.json
 # find the training folder
-LATEST_FOLDER=$(ls $BASEDIR/train_outputs/| sort | tail -1)
+LATEST_FOLDER=$(ls $BASEDIR/../train_outputs/| sort | tail -1)
 echo $LATEST_FOLDER
 # continue the previous training
-CUDA_VISIBLE_DEVICES=""  python TTS/bin/train_tacotron.py --continue_path $BASEDIR/train_outputs/$LATEST_FOLDER
+CUDA_VISIBLE_DEVICES=""  python TTS/bin/train_tacotron.py --continue_path $BASEDIR/../train_outputs/$LATEST_FOLDER
 # remove all the outputs
-rm -rf $BASEDIR/train_outputs/
+rm -rf $BASEDIR/../train_outputs/
 

tests/bash_tests/test_vocoder_gan_train.sh

@@ -3,13 +3,13 @@ set -xe
 BASEDIR=$(dirname "$0")
 echo "$BASEDIR"
 # create run dir
-mkdir $BASEDIR/train_outputs
+mkdir $BASEDIR/../train_outputs
 # run training
-CUDA_VISIBLE_DEVICES="" python TTS/bin/train_vocoder_gan.py --config_path $BASEDIR/inputs/test_vocoder_multiband_melgan_config.json
+CUDA_VISIBLE_DEVICES="" python TTS/bin/train_vocoder_gan.py --config_path $BASEDIR/../inputs/test_vocoder_multiband_melgan_config.json
 # find the training folder
-LATEST_FOLDER=$(ls $BASEDIR/train_outputs/| sort | tail -1)
+LATEST_FOLDER=$(ls $BASEDIR/../train_outputs/| sort | tail -1)
 echo $LATEST_FOLDER
 # continue the previous training
-CUDA_VISIBLE_DEVICES=""  python TTS/bin/train_vocoder_gan.py --continue_path $BASEDIR/train_outputs/$LATEST_FOLDER
+CUDA_VISIBLE_DEVICES=""  python TTS/bin/train_vocoder_gan.py --continue_path $BASEDIR/../train_outputs/$LATEST_FOLDER
 # remove all the outputs
-rm -rf $BASEDIR/train_outputs/$LATEST_FOLDER
+rm -rf $BASEDIR/../train_outputs/$LATEST_FOLDER

tests/bash_tests/test_vocoder_wavegrad_train.sh

@@ -3,13 +3,13 @@ set -xe
 BASEDIR=$(dirname "$0")
 echo "$BASEDIR"
 # create run dir
-mkdir -p $BASEDIR/train_outputs
+mkdir -p $BASEDIR/../train_outputs
 # run training
-CUDA_VISIBLE_DEVICES="" python TTS/bin/train_vocoder_wavegrad.py --config_path $BASEDIR/inputs/test_vocoder_wavegrad.json
+CUDA_VISIBLE_DEVICES="" python TTS/bin/train_vocoder_wavegrad.py --config_path $BASEDIR/../inputs/test_vocoder_wavegrad.json
 # find the training folder
-LATEST_FOLDER=$(ls $BASEDIR/train_outputs/| sort | tail -1)
+LATEST_FOLDER=$(ls $BASEDIR/../train_outputs/| sort | tail -1)
 echo $LATEST_FOLDER
 # continue the previous training
-CUDA_VISIBLE_DEVICES=""  python TTS/bin/train_vocoder_wavegrad.py --continue_path $BASEDIR/train_outputs/$LATEST_FOLDER
+CUDA_VISIBLE_DEVICES=""  python TTS/bin/train_vocoder_wavegrad.py --continue_path $BASEDIR/../train_outputs/$LATEST_FOLDER
 # remove all the outputs
-rm -rf $BASEDIR/train_outputs/$LATEST_FOLDER
+rm -rf $BASEDIR/../train_outputs/$LATEST_FOLDER

tests/bash_tests/test_vocoder_wavernn_train.sh

@@ -3,13 +3,13 @@ set -xe
 BASEDIR=$(dirname "$0")
 echo "$BASEDIR"
 # create run dir
-mkdir -p $BASEDIR/train_outputs
+mkdir -p $BASEDIR/../train_outputs
 # run training
-CUDA_VISIBLE_DEVICES="" python TTS/bin/train_vocoder_wavernn.py --config_path $BASEDIR/inputs/test_vocoder_wavernn_config.json
+CUDA_VISIBLE_DEVICES="" python TTS/bin/train_vocoder_wavernn.py --config_path $BASEDIR/../inputs/test_vocoder_wavernn_config.json
 # find the training folder
-LATEST_FOLDER=$(ls $BASEDIR/train_outputs/| sort | tail -1)
+LATEST_FOLDER=$(ls $BASEDIR/../train_outputs/| sort | tail -1)
 echo $LATEST_FOLDER
 # continue the previous training
-CUDA_VISIBLE_DEVICES=""  python TTS/bin/train_vocoder_wavernn.py --continue_path $BASEDIR/train_outputs/$LATEST_FOLDER
+CUDA_VISIBLE_DEVICES=""  python TTS/bin/train_vocoder_wavernn.py --continue_path $BASEDIR/../train_outputs/$LATEST_FOLDER
 # remove all the outputs
-rm -rf $BASEDIR/train_outputs/$LATEST_FOLDER
+rm -rf $BASEDIR/../train_outputs/$LATEST_FOLDER

tests/inputs/test_align_tts.json

@@ -65,7 +65,7 @@
     // MODEL PARAMETERS
     "positional_encoding": true,
     "hidden_channels": 256,
-    "hidden_channels_dp": 128,
+    "hidden_channels_dp": 256,
     "encoder_type": "fftransformer",
     "encoder_params":{
         "hidden_channels_ffn": 1024 ,
@@ -87,7 +87,7 @@
     "eval_batch_size":1,
     "r": 1,                 // Number of decoder frames to predict per iteration. Set the initial values if gradual training is enabled.
     "loss_masking": true,   // enable / disable loss masking against the sequence padding.
-    "phase_start_steps": [0, 40000, 80000, 160000, 170000],
+    "phase_start_steps": null,
 
 
     // LOSS PARAMETERS

tests/test_audio.py

@@ -16,7 +16,7 @@ conf = load_config(os.path.join(get_tests_input_path(), "test_config.json"))
 # pylint: disable=protected-access
 class TestAudio(unittest.TestCase):
     def __init__(self, *args, **kwargs):
-        super(TestAudio, self).__init__(*args, **kwargs)
+        super().__init__(*args, **kwargs)
         self.ap = AudioProcessor(**conf.audio)
 
     def test_audio_synthesis(self):

tests/test_loader.py

@@ -28,7 +28,7 @@ print(" > Dynamic data loader test: {}".format(DATA_EXIST))
 
 class TestTTSDataset(unittest.TestCase):
     def __init__(self, *args, **kwargs):
-        super(TestTTSDataset, self).__init__(*args, **kwargs)
+        super().__init__(*args, **kwargs)
         self.max_loader_iter = 4
         self.ap = AudioProcessor(**c.audio)