This is a repo of all the various scripts and tools we are creating or forking within the group.
| Title | Description | Link |
|---|---|---|
| NSynth Guitar Statistics | Basic stats of the guitar subset of the NSynth dataset | Link |
| Create Guitar Subset | Extract the guitar subset of the NSynth dataset | Link |
| Audio Annotator | Tool to tag or label audio samples | Link |
| Training DDSP | Script to train DDSP | Link |
| Download Preprocessed | NSynth Guitar Subset Preprocessed in tfrecord format |
Link |
| Training DDSP starting with preprocessed dataset | Script to train DDSP starting with preprocessed dataset | Link |
| Pretrained DDSP AE Checkpoint | Checkpoint for model trained on 30000 steps and 4.33 total loss | Link |
| Prepare Partial and Complete TFRecords | Complete dataset for experimenting with mapping networks or heuristics | Link |
| Audio Annotator Web | Flask app to tag or label audio samples | Link |
| Mapping Models | Installable Mapping Models package for training and inference | Link |
| Sound Generator | End to end inference pipeline for sound generation | Link |