Add abstract, prune requirements

Author: janvainer

README.md

@@ -1,6 +1,18 @@
-### Audio samples 
+## SpeedySpeech
 
-Audio samples can be found [here](https://janvainer.github.io/speedyspeech/)
+While recent neural sequence-to-sequence models have greatly improved the quality of speech synthesis, 
+there has not been a system capable of 
+fast training, fast inference and
+high-quality audio synthesis at the same time. 
+We propose a student-teacher network 
+capable of high-quality faster-than-real-time spectrogram synthesis, with low requirements on computational resources and fast training time.
+We show that self-attention layers are not necessary for generation of high quality audio. 
+We utilize simple convolutional blocks with residual connections in both student and teacher networks and use only a single attention layer in the teacher model.
+Coupled with a MelGAN vocoder, our model's voice quality was rated significantly higher than Tacotron~2.
+Our model can be efficiently trained on a single GPU and can run in real time even on a 
+CPU.
+
+See our audio samples [here](https://janvainer.github.io/speedyspeech/).
 
 ### Installation instructions
 The code was tested with `python 3.7.3` and `cuda 10.0.130`.

requirements.txt

@@ -1,76 +1,13 @@
-absl-py==0.7.1
-astor==0.8.0
-atomicwrites==1.3.0
-attrs==19.1.0
-audioread==2.1.8
-beautifulsoup4==4.8.0
-cffi==1.12.3
-clldutils==3.1.1
-colorlog==4.0.2
-csvw==1.6.0
-cycler==0.10.0
-decorator==4.4.0
-fastdtw==0.3.4
-future==0.18.0
-gast==0.2.2
-google-pasta==0.1.7
-grpcio==1.23.0
-h5py==2.9.0
-importlib-metadata==0.20
-isodate==0.6.0
-joblib==0.13.2
-kiwisolver==1.1.0
-librosa==0.7.0
-llvmlite==0.29.0
-Markdown==3.1.1
-matplotlib==3.1.1
-more-itertools==7.2.0
-numba==0.45.1
-numpy==1.16.4
-packaging==19.1
-pandas==1.0.1
-Pillow==6.1.0
-pkg-resources==0.0.0
-playsound==1.2.2
-pluggy==0.12.0
-protobuf==3.9.1
-py==1.8.0
-pycparser==2.19
-pydeps==1.7.3
-pyparsing==2.4.2
-pytest==5.1.2
-python-dateutil==2.8.0
-pytorch-ignite==0.3.0
-pytorch-lightning==0.6.0
-pytz==2019.3
-PyYAML==5.3
-regex==2019.8.19
-resampy==0.2.2
-rfc3986==1.3.2
-scikit-learn==0.21.3
-scipy==1.3.1
-segments==2.1.0
-six==1.12.0
-sklearn==0.0
-SoundFile==0.10.2
-soupsieve==1.9.3
-stdlib-list==0.5.0
-tabulate==0.8.5
-tb-nightly==1.14.0a20190603
-tensorboard==2.0.0
-termcolor==1.1.0
-torch==1.2.0
-torch-stft==0.1.4
-torchaudio==0.3.0
-torchcontrib==0.0.2
-torchnmf==0.2
-torchvision==0.4.0
-tqdm==4.42.1
-uritemplate==3.0.0
-wcwidth==0.1.7
-Werkzeug==0.15.5
-wrapt==1.11.2
-zipp==0.6.0
-barbar==0.2.1
-gitdb==4.0.2
-g2p-en==2.1.0
+torch
+barbar
+librosa
+g2p-en
+gitdb
+numpy
+fastdtw
+tensorboard
+matplotlib
+scipy
+tqdm
+pillow
+pyyaml