using waveglow, pillow, text changes

.gitmodules

@@ -0,0 +1,3 @@
+[submodule "waveglow"]
+	path = waveglow
+	url = https://github.com/NVIDIA/waveglow

README.md

@@ -1,6 +1,6 @@
 # Tacotron 2 (without wavenet)
 
-PyTorch implementation of [Natural TTS Synthesis By Conditioning
+Tacotron 2 PyTorch implementation of [Natural TTS Synthesis By Conditioning
 Wavenet On Mel Spectrogram Predictions](https://arxiv.org/pdf/1712.05884.pdf). 
 
 This implementation includes **distributed** and **fp16** support
@@ -11,7 +11,9 @@ Distributed and FP16 support relies on work by Christian Sarofeen and NVIDIA's
 
 ![Alignment, Predicted Mel Spectrogram, Target Mel Spectrogram](tensorboard.png)
 
-Visit our [website] for audio samples.
+[Download demo audio](https://github.com/NVIDIA/tacotron2/blob/master/demo.wav) trained on LJS and using Ryuchi Yamamoto's [pre-trained Mixture of Logistics
+wavenet](https://github.com/r9y9/wavenet_vocoder/)  
+"Scientists at the CERN laboratory say they have discovered a new particle."
 
 ## Pre-requisites
 1. NVIDIA GPU + CUDA cuDNN
@@ -22,9 +24,11 @@ Visit our [website] for audio samples.
 3. CD into this repo: `cd tacotron2`
 4. Update .wav paths: `sed -i -- 's,DUMMY,ljs_dataset_folder/wavs,g' filelists/*.txt`
     - Alternatively, set `load_mel_from_disk=True` in `hparams.py` and update mel-spectrogram paths 
-5. Install [PyTorch 1.0]
+5. Install [pytorch 0.4](https://github.com/pytorch/pytorch)
 6. Install python requirements or build docker image 
     - Install python requirements: `pip install -r requirements.txt`
+    - **OR**
+    - Build docker image: `docker build --tag tacotron2 .` 
 
 ## Training
 1. `python train.py --output_directory=outdir --log_directory=logdir`
@@ -33,22 +37,17 @@ Visit our [website] for audio samples.
 ## Multi-GPU (distributed) and FP16 Training
 1. `python -m multiproc train.py --output_directory=outdir --log_directory=logdir --hparams=distributed_run=True,fp16_run=True`
 
-## Inference demo
-1. Download our published [Tacotron 2] model
-2. Download our published [WaveGlow] model
-3. `jupyter notebook --ip=127.0.0.1 --port=31337`
-4. Load inference.ipynb 
+## Inference
+When performing Mel-Spectrogram to Audio synthesis with a WaveNet model, make sure Tacotron 2 and WaveNet were trained on the same mel-spectrogram representation. Follow these steps to use a a simple inference pipeline using griffin-lim:
+
+1. `jupyter notebook --ip=127.0.0.1 --port=31337`
+2. load inference.ipynb 
 
-N.b.  When performing Mel-Spectrogram to Audio synthesis, make sure Tacotron 2
-and the Mel decoder were trained on the same mel-spectrogram representation. 
 
 
 ## Related repos
-[WaveGlow](https://github.com/NVIDIA/WaveGlow) Faster than real time Flow-based
-Generative Network for Speech Synthesis
-
-[nv-wavenet](https://github.com/NVIDIA/nv-wavenet/) Faster than real time
-WaveNet.
+[nv-wavenet](https://github.com/NVIDIA/nv-wavenet/): Faster than real-time
+wavenet inference
 
 ## Acknowledgements
 This implementation uses code from the following repos: [Keith
@@ -62,7 +61,3 @@ We are thankful to the Tacotron 2 paper authors, specially Jonathan Shen, Yuxuan
 Wang and Zongheng Yang.
 
 
-[WaveGlow]: https://drive.google.com/file/d/1cjKPHbtAMh_4HTHmuIGNkbOkPBD9qwhj/view?usp=sharing
-[Tacotron 2]: https://drive.google.com/file/d/1c5ZTuT7J08wLUoVZ2KkUs_VdZuJ86ZqA/view?usp=sharing
-[pytorch 1.0]: https://github.com/pytorch/pytorch#installation
-[website]: https://nv-adlr.github.io/WaveGlow

distributed.py

@@ -164,9 +164,9 @@ def apply_gradient_allreduce(module):
                 param._execution_engine.queue_callback(allreduce_params)
             if param.requires_grad:
                 param.register_hook(allreduce_hook)
-   
+
         def set_needs_reduction(self, input, output):
             self.needs_reduction = True
-        
+
         module.register_forward_hook(set_needs_reduction)
         return module

requirements.txt

@@ -1,3 +1,4 @@
+torch==0.4.0
 matplotlib==2.1.0
 tensorflow
 numpy==1.13.3