Add pipeline tag and paper link
#3
by nielsr HF Staff - opened
README.md
CHANGED
|
@@ -1,9 +1,10 @@
|
|
| 1 |
---
|
| 2 |
library_name: KVAE-Audio
|
|
|
|
|
|
|
| 3 |
tags:
|
| 4 |
- vae
|
| 5 |
- audio
|
| 6 |
-
license: mit
|
| 7 |
---
|
| 8 |
|
| 9 |
<div align="center">
|
|
@@ -11,7 +12,7 @@ license: mit
|
|
| 11 |
<img src="assets/kvae_audio.png">
|
| 12 |
</picture>
|
| 13 |
|
| 14 |
-
<a href="https://github.com/kandinskylab/kvae-audio">GitHub</a> | <a href="https://habr.com/ru/companies/sberbank/articles/1053410/">Habr article</a> | <a href="https://kandinskylab.ai/">Project Page</a> | Technical Report (soon)
|
| 15 |
</div>
|
| 16 |
|
| 17 |
<h1>KVAE-Audio</h1>
|
|
@@ -32,30 +33,30 @@ Generative quality is established under a **fixed generator** — same DiT archi
|
|
| 32 |
|
| 33 |
### AudioCaps test set
|
| 34 |
|
| 35 |
-
| Model
|
| 36 |
| --------------- | -------- | ---------- | --------- | --------- | --------- | ------------ | ------------ | ------------- |
|
| 37 |
-
| MMAudio 44.1kHz | 427.6M
|
| 38 |
-
| DACVAE MovieGen | 107.7M
|
| 39 |
-
| SAME-L
|
| 40 |
-
| KVAE-Audio
|
| 41 |
|
| 42 |
#### Song Describer
|
| 43 |
|
| 44 |
-
| Model
|
| 45 |
| --------------- | -------- | ---------- | --------- | --------- | --------- | ------------ | ------------ | ------------- |
|
| 46 |
-
| MMAudio 44.1kHz | 427.6M
|
| 47 |
-
| DACVAE MovieGen | 107.7M
|
| 48 |
-
| SAME-L
|
| 49 |
-
| KVAE-Audio
|
| 50 |
|
| 51 |
#### LibriSpeech test-clean
|
| 52 |
|
| 53 |
-
| Model
|
| 54 |
| --------------- | -------- | ---------- | --------- | --------- | --------- | ------------ | ------------ | ------------- | --------- | --------- |
|
| 55 |
-
| MMAudio 44.1kHz | 427.6M
|
| 56 |
-
| DACVAE MovieGen | 107.7M
|
| 57 |
-
| SAME-L
|
| 58 |
-
| KVAE-Audio
|
| 59 |
|
| 60 |
### Reconstructions
|
| 61 |
|
|
@@ -63,27 +64,27 @@ Reconstruction is evaluated on open datasets across domains (the released weight
|
|
| 63 |
|
| 64 |
#### AudioSet eval
|
| 65 |
|
| 66 |
-
| Model
|
| 67 |
| --------------- | -------- | ---------- | --------- | --------- | --------- | --------- | ---------- | ---------- |
|
| 68 |
-
| MMAudio 44.1kHz | 427.6M
|
| 69 |
-
| DACVAE MovieGen | 107.7M
|
| 70 |
-
| SAME-L
|
| 71 |
-
| KVAE-Audio
|
| 72 |
|
| 73 |
#### MUSDB18-HQ
|
| 74 |
|
| 75 |
-
| Model
|
| 76 |
| --------------- | -------- | ---------- | --------- | --------- | --------- | ---------- | ---------- | ---------- |
|
| 77 |
-
| MMAudio 44.1kHz | 427.6M
|
| 78 |
-
| DACVAE MovieGen | 107.7M
|
| 79 |
-
| SAME-L
|
| 80 |
-
| KVAE-Audio
|
| 81 |
|
| 82 |
#### EARS
|
| 83 |
|
| 84 |
-
| Model
|
| 85 |
| --------------- | -------- | ---------- | --------- | --------- | --------- | ---------- | ---------- | ---------- | --------- |
|
| 86 |
-
| MMAudio 44.1kHz | 427.6M
|
| 87 |
-
| DACVAE MovieGen | 107.7M
|
| 88 |
-
| SAME-L
|
| 89 |
-
| KVAE-Audio
|
|
|
|
| 1 |
---
|
| 2 |
library_name: KVAE-Audio
|
| 3 |
+
license: mit
|
| 4 |
+
pipeline_tag: audio-to-audio
|
| 5 |
tags:
|
| 6 |
- vae
|
| 7 |
- audio
|
|
|
|
| 8 |
---
|
| 9 |
|
| 10 |
<div align="center">
|
|
|
|
| 12 |
<img src="assets/kvae_audio.png">
|
| 13 |
</picture>
|
| 14 |
|
| 15 |
+
<a href="https://github.com/kandinskylab/kvae-audio">GitHub</a> | <a href="https://habr.com/ru/companies/sberbank/articles/1053410/">Habr article</a> | <a href="https://kandinskylab.ai/">Project Page</a> | <a href="https://huggingface.co/papers/2608.05798">Paper</a> | Technical Report (soon)
|
| 16 |
</div>
|
| 17 |
|
| 18 |
<h1>KVAE-Audio</h1>
|
|
|
|
| 33 |
|
| 34 |
### AudioCaps test set
|
| 35 |
|
| 36 |
+
| Model | # Params | Latent dim | CLAP↑ | CE↑ | PQ↑ | FAD (PANNs)↓ | FAD (PASST)↓ | FAD (VGGIsh)↓ |
|
| 37 |
| --------------- | -------- | ---------- | --------- | --------- | --------- | ------------ | ------------ | ------------- |
|
| 38 |
+
| MMAudio 44.1kHz | 427.6M | 40 | *0,336* | *3,909* | *6,192* | *17,873* | *195,910* | 1,364 |
|
| 39 |
+
| DACVAE MovieGen | 107.7M | 128 | 0,313 | 3,772 | 6,167 | 20,558 | 234,312 | 1,700 |
|
| 40 |
+
| SAME-L | 852.1M | 256 | 0,322 | 3,588 | 5,756 | 18,446 | 240,635 | *1,325* |
|
| 41 |
+
| KVAE-Audio | 166.9M | 64 | **0,344** | **3,982** | **6,242** | **15,381** | **193,760** | **1,210** |
|
| 42 |
|
| 43 |
#### Song Describer
|
| 44 |
|
| 45 |
+
| Model | # Params | Latent dim | CLAP↑ | CE↑ | PQ↑ | FAD (PANNs)↓ | FAD (PASST)↓ | FAD (VGGIsh)↓ |
|
| 46 |
| --------------- | -------- | ---------- | --------- | --------- | --------- | ------------ | ------------ | ------------- |
|
| 47 |
+
| MMAudio 44.1kHz | 427.6M | 40 | **0,356** | *7,136* | *7,707* | **5,412** | **158,599** | **0,356** |
|
| 48 |
+
| DACVAE MovieGen | 107.7M | 128 | 0,312 | 6,953 | 7,538 | 10,194 | 214,009 | 1,046 |
|
| 49 |
+
| SAME-L | 852.1M | 256 | *0,345* | 7,076 | 7,465 | 8,442 | 250,668 | 0,987 |
|
| 50 |
+
| KVAE-Audio | 166.9M | 64 | 0,339 | **7,216** | **7,929** | *7,971* | *189,427* | *0,599* |
|
| 51 |
|
| 52 |
#### LibriSpeech test-clean
|
| 53 |
|
| 54 |
+
| Model | # Params | Latent dim | CLAP↑ | CE↑ | PQ↑ | FAD (PANNs)↓ | FAD (PASST)↓ | FAD (VGGIsh)↓ | WER↓ | CER↓ |
|
| 55 |
| --------------- | -------- | ---------- | --------- | --------- | --------- | ------------ | ------------ | ------------- | --------- | --------- |
|
| 56 |
+
| MMAudio 44.1kHz | 427.6M | 40 | 0,368 | *5,704* | 6,629 | 8,305 | **105,931** | *2,001* | *0,257* | *0,593* |
|
| 57 |
+
| DACVAE MovieGen | 107.7M | 128 | **0,413** | 5,482 | **7,052** | *5,008* | 210,478 | **1,501** | 0,911 | 1,048 |
|
| 58 |
+
| SAME-L | 852.1M | 256 | 0,379 | 4,617 | 5,024 | 10,257 | 301,508 | 2,721 | 0,349 | 0,629 |
|
| 59 |
+
| KVAE-Audio | 166.9M | 64 | *0,389* | **5,906** | *6,940* | **4,677** | *185,609* | 2,138 | **0,244** | **0,576** |
|
| 60 |
|
| 61 |
### Reconstructions
|
| 62 |
|
|
|
|
| 64 |
|
| 65 |
#### AudioSet eval
|
| 66 |
|
| 67 |
+
| Model | # Params | Latent dim | MEL↓ | STFT↓ | Waveform↓ | SI-SDR↑ | SDR↑ | SNR↑ |
|
| 68 |
| --------------- | -------- | ---------- | --------- | --------- | --------- | --------- | ---------- | ---------- |
|
| 69 |
+
| MMAudio 44.1kHz | 427.6M | 40 | *0,636* | *1,938* | 0,106 | -32,080 | -2,682 | -2,686 |
|
| 70 |
+
| DACVAE MovieGen | 107.7M | 128 | 0,669 | 2,275 | 0,029 | 8,384 | 9,421 | 9,416 |
|
| 71 |
+
| SAME-L | 852.1M | 256 | 0,986 | 2,726 | *0,027* | **9,586** | **10,347** | **10,339** |
|
| 72 |
+
| KVAE-Audio | 166.9M | 64 | **0,537** | **1,770** | **0,027** | *9,065* | *9,920* | *9,933* |
|
| 73 |
|
| 74 |
#### MUSDB18-HQ
|
| 75 |
|
| 76 |
+
| Model | # Params | Latent dim | MEL↓ | STFT↓ | Waveform↓ | SI-SDR↑ | SDR↑ | SNR↑ |
|
| 77 |
| --------------- | -------- | ---------- | --------- | --------- | --------- | ---------- | ---------- | ---------- |
|
| 78 |
+
| MMAudio 44.1kHz | 427.6M | 40 | 0,681 | 1,865 | 0,114 | -40,204 | -3,274 | -3,273 |
|
| 79 |
+
| DACVAE MovieGen | 107.7M | 128 | *0,519* | *1,762* | 0,024 | 9,688 | 10,046 | 10,047 |
|
| 80 |
+
| SAME-L | 852.1M | 256 | 0,668 | 1,786 | *0,023* | *10,278* | *10,648* | *10,648* |
|
| 81 |
+
| KVAE-Audio | 166.9M | 64 | **0,516** | **1,725** | **0,022** | **10,390** | **10,675** | **10,677** |
|
| 82 |
|
| 83 |
#### EARS
|
| 84 |
|
| 85 |
+
| Model | # Params | Latent dim | MEL↓ | STFT↓ | Waveform↓ | SI-SDR↑ | SDR↑ | SNR↑ | PESQ↑ |
|
| 86 |
| --------------- | -------- | ---------- | --------- | --------- | --------- | ---------- | ---------- | ---------- | --------- |
|
| 87 |
+
| MMAudio 44.1kHz | 427.6M | 40 | 0,616 | 1,395 | 0,030 | -29,947 | -2,728 | -2,697 | 2,424 |
|
| 88 |
+
| DACVAE MovieGen | 107.7M | 128 | **0,453** | **1,310** | *0,006* | **10,264** | **10,680** | **10,681** | *4,246* |
|
| 89 |
+
| SAME-L | 852.1M | 256 | 0,774 | 1,575 | 0,007 | 9,939 | 10,374 | 10,376 | 2,982 |
|
| 90 |
+
| KVAE-Audio | 166.9M | 64 | *0,463* | *1,314* | **0,006** | *9,952* | *10,377* | *10,384* | **4,266** |
|