harmonai/dance-diffusion 🔢❓ → 🖼️

▶️ 5.2K runs 📅 Sep 2022 ⚙️ Cog 0.4.3 🔗 GitHub 📄 Paper ⚖️ License
music-generation sound-effect-generation

About

Tools to train a generative model on arbitrary audio samples

Example Output

Output

Example output

Performance Metrics

137.00s Prediction Time
475.17s Total Time
All Input Parameters 
{
  "steps": 100,
  "length": 16,
  "batch_size": 1,
  "model_name": "maestro-150k"
}
Input Parameters
steps Type: integerDefault: 100
Number of steps, higher numbers will give more refined output but will take longer. The maximum is 150.
length Type: numberDefault: 8
Number of seconds to generate
batch_size Type: integerDefault: 1
How many samples to generate
model_name Default: maestro-150k
Model
Output Schema

Output

Type: string • Format: uri

Example Execution Logs 
sample_size 253952
Creating the model...
Model created
All samples
sample #1
torch.Size([2, 253952])
Skipping section, uncheck 'skip_for_run_all' to enable
Version Details
Version ID
7f27a989b169da7e984136b1185b294b744f4c38090f4a91c13de5b34a09ee7e
Version Created
September 29, 2022
Run on Replicate →