Instrument Recognition on NSynth
Accuracy
- 2022-07-12 — melspect: Accuracy 72.1
- 2025-02-17 — MATPAC (SSL, linear eval): Accuracy 74.6
- 2025-03-28 — M2D-CLAP: Accuracy 80.6
| Rank | Model | Accuracy | Paper | Code | Year |
|---|---|---|---|---|---|
| 1 | M2D-CLAP | 80.6 | M2D2: Exploring General-purpose Audio-Language Representations Beyond CLAP | nttcslab/m2d · nttcslab/eval-audio-repr | 2025 |
| 2 | M2D2 AS+ | 79.7 | M2D2: Exploring General-purpose Audio-Language Representations Beyond CLAP | nttcslab/m2d · nttcslab/eval-audio-repr | 2025 |
| 3 | M2D AS | 78.7 | M2D2: Exploring General-purpose Audio-Language Representations Beyond CLAP | nttcslab/m2d · nttcslab/eval-audio-repr | 2025 |
| 4 | MATPAC (SSL, linear eval) | 74.6 | Masked Latent Prediction and Classification for Self-Supervised Audio Representation Learning | aurianworld/matpac | 2025 |
| 5 | melspect | 72.1 | EfficientLEAF: A Faster LEarnable Audio Frontend of Questionable Use | cpjku/efficientleaf | 2022 |
| 6 | EfficientLEAF | 71.7 | EfficientLEAF: A Faster LEarnable Audio Frontend of Questionable Use | cpjku/efficientleaf | 2022 |
| 7 | LEAF | 69.2 | EfficientLEAF: A Faster LEarnable Audio Frontend of Questionable Use | cpjku/efficientleaf | 2022 |