MetricGAN (code, colab, models, paper)

.gitattributes

@@ -33,3 +33,8 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+CGA-MGAN.[[:space:]]Metric[[:space:]]GAN[[:space:]]Based[[:space:]]on[[:space:]]Convolution-Augmented[[:space:]]Gated[[:space:]]Attention[[:space:]]for[[:space:]]Speech[[:space:]]Enhancement.pdf filter=lfs diff=lfs merge=lfs -text
+MetricGAN-OKD.[[:space:]]Multi-Metric[[:space:]]Optimization[[:space:]]of[[:space:]]MetricGAN[[:space:]]via[[:space:]]Online[[:space:]]Knowledge[[:space:]]Distillation[[:space:]]for[[:space:]]Speech[[:space:]]Enhancement.pdf filter=lfs diff=lfs merge=lfs -text
+MetricGAN.[[:space:]]Generative[[:space:]]Adversarial[[:space:]]Networks[[:space:]]based[[:space:]]Black-box[[:space:]]Metric[[:space:]]Scores[[:space:]]Optimization[[:space:]]for[[:space:]]Speech[[:space:]]Enhancement.pdf filter=lfs diff=lfs merge=lfs -text
+MetricGAN+.[[:space:]]An[[:space:]]Improved[[:space:]]Version[[:space:]]of[[:space:]]MetricGAN[[:space:]]for[[:space:]]Speech[[:space:]]Enhancement.pdf filter=lfs diff=lfs merge=lfs -text
+models/metricgan-plus-voicebank/example.wav filter=lfs diff=lfs merge=lfs -text

CGA-MGAN. Metric GAN Based on Convolution-Augmented Gated Attention for Speech Enhancement.pdf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c52c8a95bc76a9e07d6f381d28e82da5fff1036df7a5f5af1d1e11b212f12b2f
+size 3329413

MetricGAN+. An Improved Version of MetricGAN for Speech Enhancement.pdf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d67b76f70f41c5892090d1b407b4731cc3d0000cfee118aa23230fad73739aad
+size 987063

MetricGAN-OKD. Multi-Metric Optimization of MetricGAN via Online Knowledge Distillation for Speech Enhancement.pdf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1dffeb81749c0ef46a39b8d380f7d21921c801549c2e8b00e1663bd46704507a
+size 2827824

MetricGAN. Generative Adversarial Networks based Black-box Metric Scores Optimization for Speech Enhancement.pdf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bb5342f6197c6e5fea3440a5fac0cb1ec955713b7ab5393707fa05df9e959dcc
+size 3092851

code/Audio_procesing.zip

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9006bb21dd71002bb18581cf42267748229dbc638c9be3d5057695e0a9c9824a
+size 362533

code/Conv-TasNet-MetricGAN-Multi-Discriminator.zip

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e4dc204a4c94637fcf5efaa21463b927b3451ffb35ba89705ff40e3c7a380937
+size 517306083

code/Dereverb_MetricGAN-U.zip

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e6b3a8d73c2f24bd66d2c1c7e0ba47b59018d79a0bd50fe4f77e529fdb32ee60
+size 315373

code/MetricGAN-KAN.zip

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:326b4a88b72067ec2a9692e3ff1c7bcf689230e9ec2935e6d7960037ea885aa7
+size 299490

code/MetricGAN-OKD [u-hyszk] +1.zip

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:56777dde4600042d83fc130da1c64f95b8600c9c2537ae2a0deb13e26115d0c8
+size 102434

code/MetricGAN-OKD.zip

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8a626c8feeeca38d93a50e66e9fe0eb17cbf1456afa6eedf0f16207439b1454c
+size 85140837

code/MetricGAN-Plus.zip

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:852cd71426ce6b3a485fec4d2aeee819efcab56c1fc376f5f363c3c131e25357
+size 14651401

code/MetricGAN-Reloaded.zip

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1abdbab5a541f727ed2521661cfa54b347dbc32801d4a254710aaa492530b717
+size 3826437521

code/MetricGAN-plus-pytorch.zip

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:70254879aadba4576007564ff18a0b6365ee6af398ff0f241f7fef7c560abd96
+size 97064073

code/MetricGAN.zip

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b6b2aeafd337cacb3ff719706e368925e8b4311e796b693810e4aae3b8e7ebec
+size 5340258

code/MetricGAN_v1.zip

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:43c01492f0d5059c5674d67020215a6387e58de8d52dbaa0f5f3ef7c6dfd30d4
+size 60974

code/SpeechBrain_MetricGAN.zip

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c304a093673a90de779a1280f10d95a546b4574e8cb568b30f0ebfc9412b1dc5
+size 33691

code/Vk_MetricGan.zip

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0179ede8b7d55a385ea08dbaf613dc0be221b9ea0177673e2093d4e0cf7cc708
+size 1711799

code/metricgan-plus-minus.zip

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:35e5de1ba2b63d68a50a172ed90159de7a7525960bcaa4d2233da4c56d9b64d9
+size 53530

code/speech-enhancement.zip

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8c5a8447a4a8ad8ff5443b7e7bb4a0d3975fef59d90dffb926dbf051220f7b84
+size 34450

models/MetricGAN-OKD/1_PE_CS_Table2.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4075b22087d30fd4dc01b96411d35d4e6a62a8ddf5bd316e9e6eb192b1553fcc
+size 26117151

models/MetricGAN-OKD/2_CS_PE_Table2.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:904af3c229f207c5f31227048e3be404191756368e872687d7246f8d1643661b
+size 26117151

models/MetricGAN-OKD/3_PE_CS+CB+CO_Table3.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8d86f7625b8bd8e2272a2b9738f67f1fdf8aeea6616e4db56931685c4d6fbf81
+size 26117343

models/MetricGAN-OKD/4_PE_CS+CB_Table4.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b0c2965f6babe681fa0af53ed3aa69b216e17d96877eface2bdc06e52b9a3b36
+size 13285890

models/MetricGAN-OKD/MetricGAN-OKD-1.weights.zip

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:45ace569ab2ccf3d24e5643d666d203416ec195755780bccf0b822dcd035def8
+size 32362

models/MetricGAN-OKD/source.txt

@@ -0,0 +1 @@
+https://github.com/wooseok-shin/MetricGAN-OKD

models/MetricGAN-Reloaded/README.md

@@ -0,0 +1,11 @@
+## Note
+
+Note that the epochX directory uses 0-indexing, while the checkpoints use 1-indexing.
+This is somewhat unfortunate, but we decided to not change it so as to not use progress
+(the new_pairs array does not depend on whether we use 0- or 1- indexing, since it is
+saved across epochs -- hence thankfully this does not affect correctness).
+
+The files in `epoch7/` are the cleaned versions of those in `data/MS-SNSD-dataset-30/train/noisy`
+produced by our model after epoch 8 (1-indexed 8).
+
+We did not include previous checkpoints in this repository to save space.

models/MetricGAN-Reloaded/checkpt-dis-8.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e2515651bb6e40b93cc4f094de7b78c77c1c37e34d0a8d5ef7955f793b102e61
+size 3754370

models/MetricGAN-Reloaded/checkpt-disopt-8.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3ca286b59d303f75e51afcbe04cfc9cdf38e809dfb080e80b7d862fc1642364f
+size 7399493

models/MetricGAN-Reloaded/checkpt-gen-8.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d32a9081edbe0c7965939279ef4aa61907798794c0723ca7a3da48a2d1911a7e
+size 7584219

models/MetricGAN-Reloaded/checkpt-genopt-8.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:84cd4459c28bb0fcc15836ebbe1cc4d4aa7af615d552ea8655d9460cdb0fec93
+size 15172973

models/MetricGAN-Reloaded/epoch7.zip

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ff252f7ffc76731bd7678d3f0a9da87c554e9af35610b20a834e5dc280e82e9c
+size 1455558142

models/MetricGAN-Reloaded/npairs_8.pkl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a08927cc94da8b7c7d2d48d41b29cb6db5e308c9e333a0e4bd2cb340738b91ac
+size 1116728

models/MetricGAN-Reloaded/source.txt

@@ -0,0 +1 @@
+https://github.com/testzer0/MetricGAN-Reloaded

models/metricgan-plus-pytorch/CSIG-GAN_trial1.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:10925ae4d094e320b2faa9bbd0a0263a1598438e0e1f5c097186167620ab1fe0
+size 26120424

models/metricgan-plus-pytorch/PESQ-GAN_trial1.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fb6f6ab151f5029a84d7c9ce699ad51dfbff0f1a2d83ac135eb089629473babc
+size 26116959

models/metricgan-plus-pytorch/PESQ-GAN_trial2.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0d9aba66256f6af982a73e38bae85b7ead3e4f857403cb7d0516316f6d99a622
+size 26116895

models/metricgan-plus-pytorch/PESQ-GAN_trial3.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:90c4b2785dcffc392a61a7fa2a9defa1ee14a58edd69d1c755b3dc15d5626ede
+size 26116959

models/metricgan-plus-pytorch/source.txt

@@ -0,0 +1 @@
+https://github.com/wooseok-shin/MetricGAN-plus-pytorch/releases/tag/v1.weights

models/metricgan-plus-voicebank/.gitattributes

@@ -0,0 +1,16 @@
+*.bin.* filter=lfs diff=lfs merge=lfs -text
+*.lfs.* filter=lfs diff=lfs merge=lfs -text
+*.bin filter=lfs diff=lfs merge=lfs -text
+*.h5 filter=lfs diff=lfs merge=lfs -text
+*.tflite filter=lfs diff=lfs merge=lfs -text
+*.tar.gz filter=lfs diff=lfs merge=lfs -text
+*.ot filter=lfs diff=lfs merge=lfs -text
+*.onnx filter=lfs diff=lfs merge=lfs -text
+*.arrow filter=lfs diff=lfs merge=lfs -text
+*.ftz filter=lfs diff=lfs merge=lfs -text
+*.joblib filter=lfs diff=lfs merge=lfs -text
+*.model filter=lfs diff=lfs merge=lfs -text
+*.msgpack filter=lfs diff=lfs merge=lfs -text
+*.pb filter=lfs diff=lfs merge=lfs -text
+*.pt filter=lfs diff=lfs merge=lfs -text
+*.pth filter=lfs diff=lfs merge=lfs -text

models/metricgan-plus-voicebank/README.md

@@ -0,0 +1,132 @@
+---
+language: "en"
+tags:
+- audio-to-audio 
+- speech-enhancement
+- PyTorch
+- speechbrain
+license: "apache-2.0"
+datasets:
+- Voicebank
+- DEMAND
+metrics:
+- PESQ
+- STOI
+inference: false
+---
+
+<iframe src="https://ghbtns.com/github-btn.html?user=speechbrain&repo=speechbrain&type=star&count=true&size=large&v=2" frameborder="0" scrolling="0" width="170" height="30" title="GitHub"></iframe>
+<br/><br/>
+
+# MetricGAN-trained model for Enhancement
+
+This repository provides all the necessary tools to perform enhancement with
+SpeechBrain. For a better experience we encourage you to learn more about
+[SpeechBrain](https://speechbrain.github.io). The model performance is:
+
+| Release | Test PESQ | Test STOI |
+|:-----------:|:-----:| :-----:|
+| 21-04-27 | 3.15 | 93.0 |
+
+## Install SpeechBrain
+
+First of all, please install SpeechBrain with the following command:
+
+```
+pip install speechbrain
+```
+
+Please notice that we encourage you to read our tutorials and learn more about
+[SpeechBrain](https://speechbrain.github.io).
+
+## Pretrained Usage
+
+To use the mimic-loss-trained model for enhancement, use the following simple code:
+
+```python
+import torch
+import torchaudio
+from speechbrain.inference.enhancement import SpectralMaskEnhancement
+
+enhance_model = SpectralMaskEnhancement.from_hparams(
+    source="speechbrain/metricgan-plus-voicebank",
+    savedir="pretrained_models/metricgan-plus-voicebank",
+)
+
+# Load and add fake batch dimension
+noisy = enhance_model.load_audio(
+    "speechbrain/metricgan-plus-voicebank/example.wav"
+).unsqueeze(0)
+
+# Add relative length tensor
+enhanced = enhance_model.enhance_batch(noisy, lengths=torch.tensor([1.]))
+
+# Saving enhanced signal on disk
+torchaudio.save('enhanced.wav', enhanced.cpu(), 16000)
+```
+
+The system is trained with recordings sampled at 16kHz (single channel).
+The code will automatically normalize your audio (i.e., resampling + mono channel selection) when calling *enhance_file* if needed. Make sure your input tensor is compliant with the expected sampling rate if you use *enhance_batch* as in the example.
+
+### Inference on GPU
+To perform inference on the GPU, add  `run_opts={"device":"cuda"}`  when calling the `from_hparams` method.
+
+### Training
+The model was trained with SpeechBrain (d0accc8).
+To train it from scratch follows these steps:
+1. Clone SpeechBrain:
+```bash
+git clone https://github.com/speechbrain/speechbrain/
+```
+2. Install it:
+```
+cd speechbrain
+pip install -r requirements.txt
+pip install -e .
+```
+
+3. Run Training:
+```
+cd  recipes/Voicebank/enhance/MetricGAN
+python train.py hparams/train.yaml --data_folder=your_data_folder
+```
+
+You can find our training results (models, logs, etc) [here](https://drive.google.com/drive/folders/1fcVP52gHgoMX9diNN1JxX_My5KaRNZWs?usp=sharing).
+
+### Limitations
+The SpeechBrain team does not provide any warranty on the performance achieved by this model when used on other datasets.
+
+## Referencing MetricGAN+
+
+If you find MetricGAN+ useful, please cite:
+
+```
+@article{fu2021metricgan+,
+  title={MetricGAN+: An Improved Version of MetricGAN for Speech Enhancement},
+  author={Fu, Szu-Wei and Yu, Cheng and Hsieh, Tsun-An and Plantinga, Peter and Ravanelli, Mirco and Lu, Xugang and Tsao, Yu},
+  journal={arXiv preprint arXiv:2104.03538},
+  year={2021}
+}
+```
+
+# **About SpeechBrain**
+- Website: https://speechbrain.github.io/
+- Code: https://github.com/speechbrain/speechbrain/
+- HuggingFace: https://huggingface.co/speechbrain/
+
+
+# **Citing SpeechBrain**
+Please, cite SpeechBrain if you use it for your research or business.
+
+
+```bibtex
+@misc{speechbrain,
+  title={{SpeechBrain}: A General-Purpose Speech Toolkit},
+  author={Mirco Ravanelli and Titouan Parcollet and Peter Plantinga and Aku Rouhe and Samuele Cornell and Loren Lugosch and Cem Subakan and Nauman Dawalatabad and Abdelwahab Heba and Jianyuan Zhong and Ju-Chieh Chou and Sung-Lin Yeh and Szu-Wei Fu and Chien-Feng Liao and Elena Rastorgueva and François Grondin and William Aris and Hwidong Na and Yan Gao and Renato De Mori and Yoshua Bengio},
+  year={2021},
+  eprint={2106.04624},
+  archivePrefix={arXiv},
+  primaryClass={eess.AS},
+  note={arXiv:2106.04624}
+}
+```

models/metricgan-plus-voicebank/config.json

@@ -0,0 +1,3 @@
+{
+  "speechbrain_interface": "SpectralMaskEnhancement"
+}

models/metricgan-plus-voicebank/enhance_model.ckpt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:147bfb866bac8264603546e035bf283370e716ed2f4b7412d308d2bcee88304f
+size 7586021

models/metricgan-plus-voicebank/example.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:72bded2034fc64a0de36250e08bd2ff3f21391e1bd941ea96b22f07e407773eb
+size 276040

models/metricgan-plus-voicebank/hyperparams.yaml

@@ -0,0 +1,40 @@
+# STFT parameters
+sample_rate: 16000
+win_length: 32
+hop_length: 16
+n_fft: 512
+window_fn: !name:torch.hamming_window
+
+compute_stft: !new:speechbrain.processing.features.STFT
+    sample_rate: !ref <sample_rate>
+    n_fft: !ref <n_fft>
+    win_length: !ref <win_length>
+    hop_length: !ref <hop_length>
+    window_fn: !ref <window_fn>
+
+compute_istft: !new:speechbrain.processing.features.ISTFT
+    sample_rate: !ref <sample_rate>
+    n_fft: !ref <n_fft>
+    win_length: !ref <win_length>
+    hop_length: !ref <hop_length>
+    window_fn: !ref <window_fn>
+
+spectral_magnitude: !name:speechbrain.processing.features.spectral_magnitude
+    power: 0.5
+
+resynth: !name:speechbrain.processing.signal_processing.resynthesize
+    stft: !ref <compute_stft>
+    istft: !ref <compute_istft>
+
+enhance_model: !new:speechbrain.lobes.models.MetricGAN.EnhancementGenerator
+    input_size: !ref <n_fft> // 2 + 1
+    hidden_size: 200
+    num_layers: 2
+    dropout: 0
+
+modules:
+    enhance_model: !ref <enhance_model>
+
+pretrainer: !new:speechbrain.utils.parameter_transfer.Pretrainer
+    loadables:
+        enhance_model: !ref <enhance_model>

models/metricgan-plus-voicebank/source.txt

@@ -0,0 +1 @@
+https://huggingface.co/speechbrain/metricgan-plus-voicebank