Anime‑XCodec2 / XCodec2 Reconstruction Demo
Compare Baseline (HKUSTAudio/xcodec2) and Fine‑tuned (NandemoGHS/Anime‑XCodec2) reconstructions side by side.
- Supported inputs: wav / flac / ogg / mp3
- Input is automatically converted to 16 kHz (as required by XCodec2).
- ZeroGPU ready. If no GPU is available, it falls back to CPU (slower).
3 60
Baseline model: HKUSTAudio/xcodec2
Fine‑tuned model: NandemoGHS/Anime-XCodec2
Inference device: auto (GPU on ZeroGPU)