Dataset Viewer
The dataset viewer should be available soon. Please retry later.
VOXPrint model index
The list of models that VOXPrint AI Media Suite apps can install. VOXPrint MDL (the suite model downloader in voxprint-common) reads it to find each model's descriptor.
Files
index.json is a JSON list. Each entry:
| Key | Meaning |
|---|---|
id |
Model id, used by the apps' model lists. |
repo |
Hugging Face model repo that holds voxprint-model.json (and our copy of every listed file). |
revision |
Commit of that repo. voxprint-model.json and every listed file exist at this commit. |
sha256 |
SHA-256 of that voxprint-model.json. |
How MDL uses it
- Fetch
index.json(falls back to the snapshot bundled invoxprint-common). - For a model id, download
https://huggingface.co/<repo>/resolve/<revision>/voxprint-model.jsonand check it againstsha256. - The descriptor (schema
voxprint-model/1) lists the files with size and SHA-256, the mirrors (Hugging Face, Kaggle, ModelScope, URL), the VRAM tier in GiB, the licence (SPDX) and flags. MDL downloads each file from any mirror, resumes and switches mirrors per file, and verifies the SHA-256. For third-party derivatives (for examplegemma4-31b,huihui-qwen36-27b) the first mirror is the original author's repo at a pinned commit; our byte-identical copy inrepois kept as the backup mirror. Files that exist only in our repo (our own quants) are served by our mirror.
The schema is schemas/voxprint-model.schema.json in voxprint-model-forge. Model Forge writes the descriptors and this index (voxprint-forge-cli index build|push|backfill).
Each model keeps its own licence, stated in its descriptor and model card. This index file itself is CC0-1.0.
- Downloads last month
- 710