Impelementation of audio processing related models
Note:
-
For those who are interested in this project, please read the Deepmind and PixelRNN posts below first. It describes basic stuff you need to know to do audio processing.
-
Then, read the following two papers.
-
Join our discussion here!
-
For those who are new to RNN, here are some resources. The Unreasonable Effectiveness of Recurrent Neural Networks, How to implement a recurrent neural network Part 1
paper1 : Wavenet
Deepmind Blogpost : Wavenet blogpost
paper2 : Generating Sequences With Recurrent Neural Networks
PixelRNN review : Magenta Review
- Yutaro Yamada
- Kshitijh Meelu
- Ethan Weinberger
- Krishnan Srinivasan
- Lincoln Swaine-Moore
- Henry Li
Status: [ not started ◼️ | completed ✅ | in progress 💬]
| Task | Status | Deadline | Assigned to |
|---|---|---|---|
| read paper1 & Deepmind Blogpost | ✅ | 9.30.16 | Everyone |
| read PixelRNN review | ✅ | 10.1.16 | Everyone |
| preprocess audio data to wavenet data | ✅ | 10.7.16 | Krishnan, Sumedh |
| setup an instance on GCP | 💬 | 10.7.16 | Kshitijh |
| build ToyModel | ✅ | 10.7.16 | Yutaro |
| postprocess wavenet data to audio data | ✅ | 10.7.16 | Ethan |
| pipeline preprocessed data into toy model | 💬 | 10.14.16 | Krishnan |
| Read/replicate CharRNN, look into text->speech generation | 💬 | 10.14.16 | Krishnan, Henry, Lincoln |