SingQ commited on
Commit
7bef3a7
·
verified ·
1 Parent(s): dbac9e7

Upload README.md with huggingface_hub

Browse files
Files changed (1) hide show
  1. README.md +16 -0
README.md ADDED
@@ -0,0 +1,16 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: other
3
+ tags: [coreml, singing, lyrics-alignment, source-separation]
4
+ ---
5
+ # SingQ models
6
+
7
+ Core ML models downloaded on first use by **SingQ**, a macOS app for singing practice.
8
+
9
+ | File | Use | Source | License |
10
+ | --- | --- | --- | --- |
11
+ | `wav2vec2-ko.mlpackage.zip` | Korean lyric alignment (CTC) | [kresnik/wav2vec2-large-xlsr-korean](https://huggingface.co/kresnik/wav2vec2-large-xlsr-korean), trained on [Zeroth-Korean](https://openslr.org/40/) (CC BY 4.0) | Apache-2.0 |
12
+ | `wav2vec2-en.mlpackage.zip` | English lyric alignment (CTC) | [facebook/wav2vec2-large-960h-lv60-self](https://huggingface.co/facebook/wav2vec2-large-960h-lv60-self) | Apache-2.0 |
13
+ | `htdemucs.mlpackage.zip` | Vocal / accompaniment separation | [Hybrid Transformer Demucs](https://github.com/facebookresearch/demucs) by Meta AI; Core ML conversion from [slaughters85j/htdemucs-coreml](https://huggingface.co/slaughters85j/htdemucs-coreml) | MIT |
14
+
15
+ The wav2vec2 models were converted to Core ML (English FP16, Korean FP32 because the XLS-R model overflows in FP16; fixed 34-second input at 16 kHz, log-softmax output every 20 ms) without retraining.
16
+ `models.json` lists each file's size and SHA-256.