# Voice Bench

A browser tool for exploring the Vox Day / Christopher Nuttall SFT packages
and generating scenes from them against the joint CPT+SFT model, including a
STYLE-swap control for testing voice conditioning across authors.

Full training report (CPT + SFT setup, results, and evaluation data):
https://claude.ai/code/artifact/d3415f56-3f9b-40de-aa70-ac0992549ced

## What's here

- `voice-bench.html` — the tool itself. It's a single self-contained file
  with no install step: open it in a browser (double-click it, or
  File > Open). It needs no server of its own — it just needs a model server
  to talk to (see below).
- `start-model.sh` — starts the model server (`llama-server`, from
  [llama.cpp](https://github.com/ggml-org/llama.cpp)) that `voice-bench.html`
  sends generation requests to.

The 208 training/holdout packages themselves are already embedded inside
`voice-bench.html` — nothing else needs downloading for browsing.

## One-time setup (per machine)

### 1. Get the model file

Download `Qwen3-30B-A3B-VoxDay-Nuttall-SFT-ep3-Q8_0.gguf` (~31GB) from
Google Drive:

https://drive.google.com/open?id=1q1w_XMp6oRmfx8xqXDV-HvDx_krVxQtD

**Note (2026-08-30): the link above still points to the ep2 file.** Production was just switched to ep3 (separability + length-calibration evidence, see `JOINT_TRAINING_CLOSEOUT_REPORT.md`); the ep3 GGUF upload and link swap are pending.

Memory needed: ~31GB for the weights plus ~3GB of KV cache at the full
32768-token context, so ~35GB total. Confirmed working on a DGX Spark;
should comfortably fit a 48GB L40 as well. If a machine has less than that,
lower `CTX_SIZE` (below) before anything else.

### 2. Build llama.cpp with CUDA support

Needs an NVIDIA GPU with the CUDA toolkit and `cmake` already installed.

```
git clone https://github.com/ggml-org/llama.cpp
cd llama.cpp
cmake -B build -DGGML_CUDA=ON
cmake --build build --config Release -j
```

This produces `llama.cpp/build/bin/llama-server`, which is what
`start-model.sh` runs.

### 3. Start the model server

```
MODEL_PATH=/path/to/Qwen3-30B-A3B-VoxDay-Nuttall-SFT-ep3-Q8_0.gguf \
LLAMA_SERVER=/path/to/llama.cpp/build/bin/llama-server \
./start-model.sh
```

Leave this running in a terminal — it's the process `voice-bench.html`
talks to. If you'd rather not set environment variables every time, edit the
defaults at the top of `start-model.sh` instead.

### 4. Open voice-bench.html

Open the file in a browser. The "Server" field in the top-right corner
defaults to `http://127.0.0.1:8200`, which is correct if the model server is
running on the same machine as the browser. If it's running on a different
machine on your network, change it to that machine's address instead, e.g.
`http://192.168.1.50:8200` — and make sure that machine's firewall allows
inbound connections on the port (`sudo ufw allow from <your-subnet> to any
port 8200 proto tcp` on Ubuntu).

## Troubleshooting

- **"SERVER UNREACHABLE"** in the page header — the model server isn't
  running yet, is still loading the model (can take a minute for a 31GB
  file), or the Server field points at the wrong address/port.
- **Out of memory when starting the server** — re-run with a smaller
  context, e.g. `CTX_SIZE=8192 ./start-model.sh`.
- **Opening the HTML file from a typed path does something odd in
  Firefox** — use File > Open (Ctrl+O) instead of typing the path into the
  address bar, or paste the full `file:///home/you/path/voice-bench.html`
  URL.