epsilver commited on
Commit
c752c75
·
verified ·
1 Parent(s): 42ce933

sync with current PMS models: RVM replaces U2-Net for bg removal; drop retired hubert_base.pt (runtime uses hubert.onnx); README attribution for face models + pitch bases

Browse files
README.md CHANGED
@@ -15,4 +15,14 @@ Curated model bundle for [Pop Maker Studio](https://github.com/verticalrectangle
15
  | wav2vec2-base-960h | [Xenova/wav2vec2-base-960h](https://huggingface.co/Xenova/wav2vec2-base-960h) | Apache 2.0 |
16
  | Moondream2 vision | [Xenova/moondream2](https://huggingface.co/Xenova/moondream2) | Apache 2.0 |
17
  | Piper voices | [rhasspy/piper-voices](https://huggingface.co/rhasspy/piper-voices) | MIT |
18
- | U²-Net human seg | [danielgatis/rembg](https://github.com/danielgatis/rembg) | MIT |
 
 
 
 
 
 
 
 
 
 
 
15
  | wav2vec2-base-960h | [Xenova/wav2vec2-base-960h](https://huggingface.co/Xenova/wav2vec2-base-960h) | Apache 2.0 |
16
  | Moondream2 vision | [Xenova/moondream2](https://huggingface.co/Xenova/moondream2) | Apache 2.0 |
17
  | Piper voices | [rhasspy/piper-voices](https://huggingface.co/rhasspy/piper-voices) | MIT |
18
+ | RobustVideoMatting (mobilenetv3) | [PeterL1n/RobustVideoMatting](https://github.com/PeterL1n/RobustVideoMatting) | GPL-3.0 |
19
+ | SCRFD-10G + 2d106det (face) | [InsightFace model zoo](https://github.com/deepinsight/insightface) | non-commercial research (see upstream) |
20
+ | RVC pitch bases | generated from [lj1995 pretrained_v2](https://huggingface.co/lj1995/VoiceConversionWebUI) f0G emb_pitch tensors | MIT (derivative) |
21
+
22
+ Layout note: the app expects these under `build/models/` — flat names for the
23
+ top-level models (`vision/decoder_model_merged_q4.onnx` → `moondream-decoder.onnx`,
24
+ `vision_encoder_q4.onnx` → `moondream-encoder.onnx`, `embed_tokens_q4.onnx` →
25
+ `embed_tokens.onnx`, `tokenizer.json` → `moondream-tokenizer.json`), `face/` and
26
+ `piper/` keep their subdirectories. `hubert.onnx` is exported from upstream
27
+ `hubert_base.pt` by the app's `export-hubert` tool (the .pt itself is not shipped
28
+ here — it is not used at runtime).
background/{u2netp_human_seg.onnx → rvm_mobilenetv3_fp32.onnx} RENAMED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:309c8469258dda742793dce0ebea8e6dd393174f89934733ecc8b14c76f4ddd8
3
- size 4574861
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:88d4531297118f595bf2fd60f6f566aec2e559393802d1f436c380f0cbbd2828
3
+ size 14975696
hubert/hubert_base.pt DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:f54b40fd2802423a5643779c4861af1e9ee9c1564dc9d32f54f20b5ffba7db96
3
- size 189507909