# Condition example assets

**Do not commit `.wav` or `.png` files to git** — Hugging Face Spaces reject binary pushes.

Upload assets to the FFASR Hub bucket instead (default repo: `treble-technologies/FFASR_Leaderboard-storage`), under the `examples/` prefix:

| Bucket path | Description |
|-------------|-------------|
| `examples/scene.png` | Treble scene screenshot |
| `examples/dry_speech_example_1.wav` | Near Field Speech example |
| `examples/high_snr_example_1.wav` | High SNR example |
| `examples/mid_snr_example_1.wav` | Mid SNR example |
| `examples/low_snr_example_1.wav` | Low SNR example |

### Upload with the Hugging Face CLI

```bash
export BUCKET=treble-technologies/FFASR_Leaderboard-storage

for f in scene.png dry_speech_example_1.wav high_snr_example_1.wav mid_snr_example_1.wav low_snr_example_1.wav; do
  huggingface-cli upload "$BUCKET" "examples/$f" "examples/$f" --repo-type dataset
done
```

For local development you can also place the same files in this folder; they are gitignored and loaded directly.

Optional env override: `FFASR_EXAMPLES_BUCKET_PREFIX` (default: `examples/`).