GAME Small — official v1.0.3 ONNX mirror
Community mirror maintained by Hikari Tsai (aaatmy), not an official openvpi repository.
The five FP32 ONNX graphs and config.json are copied without modification from
openvpi/GAME v1.0.3.
These are the official ONNX exports of the GAME 1.0 Small pretrained weights,
not a newly trained model or a quantized conversion.
- Hugging Face: https://huggingface.co/aaatmy/game-small-onnx
- Git mirror: https://github.com/Hikari-Tsai/music-detection/tree/staging/assets/models/game/1.0.3-small
- ONNX files: 51,054,395 bytes total (51.1 MB), opset 17.
- Source ZIP:
GAME-1.0.3-small-onnx.zip, 45,676,364 bytes. - Source ZIP SHA-256:
00ba0c64115b6b874d9ea4afd3e6cf822abda2a04e52569233b0a044fd40e4e8.
Files and usage
| File | Purpose |
|---|---|
encoder.onnx |
Encode the audio waveform into model features |
segmenter.onnx |
Iteratively predict note boundaries |
estimator.onnx |
Estimate note pitches and presence scores |
dur2bd.onnx |
Convert known durations to boundaries |
bd2dur.onnx |
Convert boundaries to durations |
config.json |
Model configuration, sample rate and language IDs |
manifest.json, SHA256SUMS |
Upstream provenance, file sizes and SHA-256 digests |
Use the graphs together as described in the versioned upstream ONNX documentation. This is not a Hugging Face Transformers or Diffusers pipeline. It expects a host implementation of preprocessing, iterative segmentation and result decoding. The configuration uses 44,100 Hz audio and a 0.01-second time step.
To download the complete bundle:
hf download aaatmy/game-small-onnx --local-dir ./game-small-onnx
For reproducible applications, add --revision <commit-sha> to pin the model snapshot.
To check the original assets after download, run shasum -a 256 -c SHA256SUMS
inside the downloaded directory (or sha256sum -c SHA256SUMS on Linux).
This mirror alone does not enable vocal-range analysis in Key & Tempo. It makes no accuracy guarantee for lead vocals, harmonies or highest/lowest notes.
Attribution and license
Model credit belongs to openvpi/GAME and its contributors; the official model release was published by yqzhishen. The upstream release credits data labels to 挚彬Club, YQ之神, 小狼 and 幽寂. See the original model release for the training sources and acknowledgements.
The model files are licensed under CC BY-NC-SA 4.0, as specified in that release. The full legal text is included in LICENSE: attribution, noncommercial use and ShareAlike conditions apply. This mirror adds only documentation and checksums; the model files and configuration are unchanged. Any future conversion or quantization must retain the applicable model terms and identify its changes.
The upstream code repository uses MIT, but its MIT license does not replace the model weights' CC BY-NC-SA 4.0 license. Likewise, the MIT license of Key & Tempo's original application code does not apply to these third-party model files. The models are provided without warranties under their license; no upstream endorsement of this mirror is implied.
繁體中文
此處託管官方 GAME Small v1.0.3 的原始 FP32 ONNX，共五個模型及設定檔， 總模型大小約 51.1 MB，未進行量化或修改。由 Hikari Tsai 建立社群副本， 並非 openvpi 官方 Hugging Face 帳戶。
模型採 CC BY-NC-SA 4.0（署名、非商業性、相同方式分享），不適用本專案的 MIT。
來源、檔案大小與校驗碼列於 manifest.json；本次僅託管模型，尚未接入網站分析流程。
- Downloads last month
- 16