Instructions to use oppiegoopie/AloudLLM with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Notebooks
Google Colab
Kaggle
Local Apps Settings

How to use oppiegoopie/AloudLLM with llama.cpp:

Install (macOS, Linux)

curl -LsSf https://llama.app/install.sh | sh
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf oppiegoopie/AloudLLM:Q8_0
# Run inference directly in the terminal:
llama cli -hf oppiegoopie/AloudLLM:Q8_0

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf oppiegoopie/AloudLLM:Q8_0
# Run inference directly in the terminal:
llama cli -hf oppiegoopie/AloudLLM:Q8_0

Build from source code

git clone https://github.com/ggml-org/llama.cpp.git
cd llama.cpp
cmake -B build
cmake --build build -j --target llama-app
# Start a local OpenAI-compatible server with a web UI:
./build/bin/llama serve -hf oppiegoopie/AloudLLM:Q8_0
# Run inference directly in the terminal:
./build/bin/llama cli -hf oppiegoopie/AloudLLM:Q8_0

Use Docker

docker model run hf.co/oppiegoopie/AloudLLM:Q8_0

Llama new
LM Studio
Jan

vLLM

How to use oppiegoopie/AloudLLM with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "oppiegoopie/AloudLLM"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "oppiegoopie/AloudLLM",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/oppiegoopie/AloudLLM:Q8_0

Ollama
How to use oppiegoopie/AloudLLM with Ollama:
```
ollama run hf.co/oppiegoopie/AloudLLM:Q8_0
```
Unsloth Desktop

How to use oppiegoopie/AloudLLM with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf oppiegoopie/AloudLLM:Q8_0

Configure the model in Pi

# Install Pi:
npm install -g @earendil-works/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "llama-cpp": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "oppiegoopie/AloudLLM:Q8_0"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use oppiegoopie/AloudLLM with Docker Model Runner:
```
docker model run hf.co/oppiegoopie/AloudLLM:Q8_0
```

Lemonade

How to use oppiegoopie/AloudLLM with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull oppiegoopie/AloudLLM:Q8_0

Run and chat with the model

lemonade run user.AloudLLM-Q8_0

List all available models

lemonade list

Hermes Agent

How to use oppiegoopie/AloudLLM with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf oppiegoopie/AloudLLM:Q8_0

Configure Hermes

# Install Hermes:
curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash
hermes setup
# Point Hermes at the local server:
hermes config set model.provider custom
hermes config set model.base_url http://127.0.0.1:8080/v1
hermes config set model.default oppiegoopie/AloudLLM:Q8_0

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use oppiegoopie/AloudLLM with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf oppiegoopie/AloudLLM:Q8_0

Configure OpenClaw

# Install OpenClaw:
npm install -g openclaw@latest
# Register the local server and set it as the default model:
openclaw onboard --non-interactive --mode local \
  --auth-choice custom-api-key \
  --custom-base-url http://127.0.0.1:8080/v1 \
  --custom-model-id "oppiegoopie/AloudLLM:Q8_0" \
  --custom-provider-id llama-cpp \
  --custom-compatibility openai \
  --custom-text-input \
  --accept-risk \
  --skip-health

Run OpenClaw

openclaw agent --local --agent main --message "Hello from Hugging Face"

AloudLLM: on-device dictation cleanup (0.8B)

A full fine-tune of Qwen3.5-0.8B that turns raw speech-recognition output into finished text for Aloud's dictation feature: punctuation, sentence boundaries, filler removal, resolved self-corrections, paragraphs for long dictations, dictionary spellings. It is trained to preserve meaning rather than rewrite, and to leave fluent speech nearly unchanged.

Built to replace a 9B model served over the network with something that runs in the app's own llama.cpp runtime: on an M4 Pro MacBook the Q8_0 file loads in under a second and cleans a typical dictation in about 0.2 s (p90 about 1 s).

Files

File	Quant	Size
`aloud-cleanup-0.8b-run2-Q8_0.gguf`	Q8_0	1.08 GB

Prompt format

The model expects Aloud's cleanup request: the Superwhisper-style system prompt with the destination and dictionary hint, the raw transcript as the user message, Qwen chat template with thinking disabled (chat_template_kwargs: {"enable_thinking": false}), temperature 0.

Training data

About 120k examples: 1.9k real dictations with reference cleanups, 99k synthetic short/medium pairs and 10k long monologue pairs generated clean-first by a 9B teacher and then noised into realistic transcripts, so every target is correct by construction. Transcripts that lost content were filtered out. Includes identity examples (already-clean input) and stutter augmentation.

Evaluation (run2, 400 frozen real dictations, blind pairwise judge vs production Qwen3.5-9B)

	Production 9B	This model
Preferred cleanup	137	102 (161 ties)
Material meaning errors	9	13
Median latency on M4 Pro	~0.7 s (network)	0.23 s

Known weakness: very long rambling dictations (500+ words) still occasionally lose a clause or flip a nuance. Aloud validates every cleanup before pasting and falls back to the spoken words when the output is implausible.

Downloads last month: 50

GGUF

Model size

1B params

Architecture

qwen35

Hardware compatibility

8-bit

Model tree for oppiegoopie/AloudLLM

Base model

Qwen/Qwen3.5-0.8B-Base

Finetuned

Qwen/Qwen3.5-0.8B

Quantized

(397)

this model