Spaces:
Runtime error
Runtime error
File size: 2,997 Bytes
4b16ff2 825c8bf 1dea888 825c8bf 65fa65c 825c8bf 65fa65c 1dea888 65fa65c 825c8bf 1dea888 825c8bf 1dea888 825c8bf 65fa65c 1dea888 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 |
# audio-diffusion
### Apply [Denoising Diffusion Probabilistic Models](https://arxiv.org/abs/2006.11239) using the new Hugging Face [diffusers](https://github.com/huggingface/diffusers) package to synthesize music instead of images.
---
Audio can be represented as images by transforming to a [mel spectrogram](https://en.wikipedia.org/wiki/Mel-frequency_cepstrum), such as the one shown above. The class `Mel` in `mel.py` can convert a slice of audio into a mel spectrogram of `x_res` x `y_res` and vice-versa. The higher the resolution, the less audio information will be lost. You can see how this works in the `test-mel.ipynb` notebook.
A DDPM model is trained on a set of mel spectrograms that have been generated from a directory of audio files. It is then used to synthesize similar mel spectrograms, which are then converted back into audio. See the `test-model.ipynb` notebook for an example.
## Generate Mel spectrogram dataset from directory of audio files
### Training can be run with Mel spectrograms of resolution 64x64 on a single commercial grade GPU (e.g. RTX 2080 Ti). The `hop_length` should be set to 1024 for better results.
```bash
python src/audio_to_images.py \
--resolution 64 \
--hop_length 1024\
--input_dir path-to-audio-files \
--output_dir data-test
```
### Generate dataset of 256x256 Mel spectrograms and push to hub (you will need to be authenticated with `huggingface-cli login`).
```bash
python src/audio_to_images.py \
--resolution 256 \
--input_dir path-to-audio-files \
--output_dir data-256 \
--push_to_hub teticio\audio-diffusion-256
```
## Train model
### Run training on local machine.
```bash
accelerate launch --config_file accelerate_local.yaml \
src/train_unconditional.py \
--dataset_name data-64 \
--resolution 64 \
--hop_length 1024 \
--output_dir ddpm-ema-audio-64 \
--train_batch_size 16 \
--num_epochs 100 \
--gradient_accumulation_steps 1 \
--learning_rate 1e-4 \
--lr_warmup_steps 500 \
--mixed_precision no
```
### Run training on local machine with `batch_size` of 1 and `gradient_accumulation_steps` 16 to compensate, so that 256x256 resolution model fits on commercial grade GPU.
```bash
accelerate launch --config_file accelerate_local.yaml \
src/train_unconditional.py \
--dataset_name teticio/audio-diffusion-256 \
--resolution 256 \
--output_dir ddpm-ema-audio-256 \
--num_epochs 100 \
--train_batch_size 1 \
--eval_batch_size 1 \
--gradient_accumulation_steps 16 \
--learning_rate 1e-4 \
--lr_warmup_steps 500 \
--mixed_precision no
```
### Run training on SageMaker.
```bash
accelerate launch --config_file accelerate_sagemaker.yaml \
src/train_unconditional.py \
--dataset_name teticio/audio-diffusion-256 \
--resolution 256 \
--output_dir ddpm-ema-audio-256 \
--train_batch_size 16 \
--num_epochs 100 \
--gradient_accumulation_steps 1 \
--learning_rate 1e-4 \
--lr_warmup_steps 500 \
--mixed_precision no
```
|