README.md

---
license: apache-2.0
library_name: vokra
tags:
  - vokra
  - gguf
  - speaker-embedding
---

# campplus-speaker-encoder (Vokra GGUF)

Speaker embedding model, converted to the Vokra GGUF format for [Vokra](https://github.com/ayutaz/vokra), a zero-dependency speech-AI inference runtime.

**This is a conversion, not a new model.** The weights are the upstream ones; Vokra re-packages them so its runtime can memory-map them directly. Credit for the model belongs upstream — see *Source* below.

## Files

| File | Size | SHA-256 |
|---|---|---|
| `campplus.gguf` | 26.4 MB | `c760971dc698fe7bfc5b9af9a4ba3b1ed1668b6c7a4e19086836b41e14285bea` |

## Usage

```bash
# Download (any HTTP client works — the file is a plain GGUF)
curl -L -o campplus.gguf \
  https://huggingface.co/vokra/campplus-speaker-encoder/resolve/main/campplus.gguf
```

```bash
# 192-d speaker embedding
vokra-cli run --model campplus.gguf --input a.wav

# Cosine similarity between two utterances
vokra-cli run --model campplus.gguf --input a.wav --compare b.wav
```

## Provenance

| Field | Value |
|---|---|
| Architecture | `campplus` |
| Tensors | 619 |
| Upstream source | iic/speech_campplus via ayousanz/campplus-onnx (Apache-2.0) |
| Upstream licence | `Apache-2.0` |
| Licence class | `permissive` |
| Registry model id | `campplus` |
| Vokra GGUF schema | 1 |
| Converted by | vokra-core 0.1.0-alpha.0 |

Every row above is read out of this file's own `vokra.*` metadata, so the card cannot claim something the artifact does not carry.

## Licence

The weights are distributed under **Apache-2.0**, unchanged from upstream. Conversion does not alter the licence, and your obligations run to the upstream author.

## Verifying this file

```bash
shasum -a 256 campplus.gguf
# expect: c760971dc698fe7bfc5b9af9a4ba3b1ed1668b6c7a4e19086836b41e14285bea
```