Add LatentSync 1.6 GGUF

.gitattributes

@@ -248,3 +248,4 @@ Voxtral-Mini-4B-Realtime-2602-GGUF/voxtral-mini-4b-realtime-2602-bf16.gguf filte
 Voxtral-Mini-4B-Realtime-2602-GGUF/voxtral-mini-4b-realtime-2602-q4_k.gguf filter=lfs diff=lfs merge=lfs -text
 Voxtral-Mini-4B-Realtime-2602-GGUF/voxtral-mini-4b-realtime-2602-q8_0.gguf filter=lfs diff=lfs merge=lfs -text
 Kokoro-82M-GGUF/kokoro-82m-multilingual-q8_0.gguf filter=lfs diff=lfs merge=lfs -text
+LatentSync-1.6-GGUF/latentsync-1.6-f32-f16.gguf filter=lfs diff=lfs merge=lfs -text

LatentSync-1.6-GGUF/README.md

@@ -0,0 +1,47 @@
+---
+license: other
+license_name: mixed-upstream-licenses
+license_link: https://github.com/deepinsight/insightface#license
+tags:
+  - audio.cpp
+  - gguf
+  - lip-sync
+base_model: ByteDance/LatentSync-1.6
+---
+
+# LatentSync 1.6 for audio.cpp
+
+Use this package with [audio.cpp](https://github.com/0xShug0/audio.cpp).
+It synchronizes lip motion in an existing video with supplied speech.
+
+> [!WARNING]
+> This local development package contains InsightFace face detection and landmark
+> weights restricted to non-commercial research. LatentSync's Apache-2.0 license
+> does not override those terms. Review upstream licenses before redistribution
+> or deployment.
+
+`latentsync-1.6-f32-f16.gguf` includes the temporal diffusion U-Net, Whisper Tiny
+audio encoder, SD VAE, SCRFD face detector, 106-point landmark model, face mask,
+configuration sidecars, and embedded audio.cpp model spec. Original mixed F32/F16
+tensor storage is retained; inference requires no Python or ONNX Runtime.
+
+Install FFmpeg on PATH, then run:
+
+```bash
+audiocpp_cli --task gen --family latentsync \
+  --model latentsync-1.6-f32-f16.gguf --backend cuda \
+  --audio speech.wav --request-option reference_video_path=speaker.mp4 \
+  --out-dir outputs/latentsync --log
+```
+
+The CLI writes RGB24 frames and a small JSON metadata sidecar. MP4 encoding and audio muxing
+are separate FFmpeg steps; see `docs/models/latentsync.md` in the port's source.
+
+## Upstream Sources
+
+- [LatentSync weights](https://huggingface.co/ByteDance/LatentSync-1.6)
+- [LatentSync code](https://github.com/bytedance/LatentSync/tree/a229c3948406bc2cf6eaf4873e662e70c6a04746)
+- [SD VAE](https://huggingface.co/stabilityai/sd-vae-ft-mse)
+- [InsightFace license](https://github.com/deepinsight/insightface#license)
+
+This is an experimental local package, not a published release.

LatentSync-1.6-GGUF/latentsync-1.6-f32-f16.gguf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3365364fc3761a7216969b4d474b4afd0c5892af01f6aa4ef05e63440714d283
+size 5448376384