Instructions to use PatrickPat/remnants-llm-v5-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Notebooks
Google Colab
Kaggle
Local Apps Settings

How to use PatrickPat/remnants-llm-v5-GGUF 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 PatrickPat/remnants-llm-v5-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf PatrickPat/remnants-llm-v5-GGUF:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf PatrickPat/remnants-llm-v5-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf PatrickPat/remnants-llm-v5-GGUF:Q4_K_M

Use pre-built binary

# Download pre-built binary from:
# https://github.com/ggerganov/llama.cpp/releases
# Start a local OpenAI-compatible server with a web UI:
./llama-server -hf PatrickPat/remnants-llm-v5-GGUF:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf PatrickPat/remnants-llm-v5-GGUF:Q4_K_M

Build from source code

git clone https://github.com/ggerganov/llama.cpp.git
cd llama.cpp
cmake -B build
cmake --build build -j --target llama-server llama-cli
# Start a local OpenAI-compatible server with a web UI:
./build/bin/llama-server -hf PatrickPat/remnants-llm-v5-GGUF:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf PatrickPat/remnants-llm-v5-GGUF:Q4_K_M

Use Docker

docker model run hf.co/PatrickPat/remnants-llm-v5-GGUF:Q4_K_M

LM Studio
Jan

vLLM

How to use PatrickPat/remnants-llm-v5-GGUF with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "PatrickPat/remnants-llm-v5-GGUF"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "PatrickPat/remnants-llm-v5-GGUF",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/PatrickPat/remnants-llm-v5-GGUF:Q4_K_M

Ollama
How to use PatrickPat/remnants-llm-v5-GGUF with Ollama:
```
ollama run hf.co/PatrickPat/remnants-llm-v5-GGUF:Q4_K_M
```
Unsloth Desktop

How to use PatrickPat/remnants-llm-v5-GGUF with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf PatrickPat/remnants-llm-v5-GGUF:Q4_K_M

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": "PatrickPat/remnants-llm-v5-GGUF:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use PatrickPat/remnants-llm-v5-GGUF with Docker Model Runner:
```
docker model run hf.co/PatrickPat/remnants-llm-v5-GGUF:Q4_K_M
```

Lemonade

How to use PatrickPat/remnants-llm-v5-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull PatrickPat/remnants-llm-v5-GGUF:Q4_K_M

Run and chat with the model

lemonade run user.remnants-llm-v5-GGUF-Q4_K_M

List all available models

lemonade list

Hermes Agent

How to use PatrickPat/remnants-llm-v5-GGUF with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf PatrickPat/remnants-llm-v5-GGUF:Q4_K_M

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 PatrickPat/remnants-llm-v5-GGUF:Q4_K_M

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use PatrickPat/remnants-llm-v5-GGUF with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf PatrickPat/remnants-llm-v5-GGUF:Q4_K_M

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 "PatrickPat/remnants-llm-v5-GGUF:Q4_K_M" \
  --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"

Remnants LLM v5 (GGUF)

A fine-tune of Qwen3-4B-Instruct-2507 for Project Remnants, a Project Zomboid mod that gives survivor NPCs a local language model.

It does two jobs:

Answers the game's order prompt with a JSON order envelope.
Speaks in character as a survivor. It knows itself as "Remnants LLM".

It is built for that mod. It was not evaluated as a general assistant.

Use with Ollama

ollama pull hf.co/PatrickPat/remnants-llm-v5-GGUF:Q4_K_M

The model name contains "remnants", so the mod sends it the short prompt it was trained on.

Files

File	Purpose
`remnants-llm-v5-Q4_K_M.gguf`	Q4_K_M weights, 2,497,280,128 bytes
`template`	Plain ChatML chat template for Ollama, without tool calling or thinking
`system`	Default system prompt for Ollama
`params`	Default sampling for Ollama: temperature 0.7, top_k 20, top_p 0.8, repeat_penalty 1, plus the two ChatML stop tokens

SHA-256 of the GGUF: aab73c9bf1e3b38abf3793284e43977ac88ea8194330c8058044f6001c0ec67b

The template file matters for Ollama. Without it, Ollama picks a thinking-style template that makes this model answer with empty text. The pulled copy was checked against the original local model at temperature 0, and the answers matched on 6 of 6 prompts, including a system prompt, a multi-turn chat and a JSON order.

Results

These come from the mod's own checks, not an independent benchmark.

Check	Result
Order lines in the game's test set	192 of 202
Fresh order lines	63 of 68
New-kind awareness cards	242 of 250
Scripted situations	20 of 24
Awareness fact check	53 of 55
Roleplay rule pass	107 of 110

Known weaknesses

It sometimes adds backstory to short lines. It did so on 12 of 45 short test lines.
It lectures on some mature-topic lines. That happened on 5 test lines.
Orders are not perfect. It missed 10 of the 202 order lines in the test set.

Content

The characters live in a zombie apocalypse. The model can use profanity and talk about violence and other dark themes. The training data was written to leave out explicit sexual content and any sexual content involving minors. There is no guarantee, so add your own filtering if you need it.

Training

A LoRA fine-tune merged into the base weights. The examples are JSON orders, in-character chat and game-fact awareness. They were written with AI assistance and checked with scripts. The training data is not published.

License

Apache-2.0, the same as the base model.

Downloads last month: 107

GGUF

Model size

4B params

Architecture

qwen3

Hardware compatibility

4-bit

Model tree for PatrickPat/remnants-llm-v5-GGUF

Base model

Qwen/Qwen3-4B-Instruct-2507

Finetuned

(2464)

this model