Instructions to use runanywhere/SmolVLM-500M-Instruct-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use runanywhere/SmolVLM-500M-Instruct-GGUF with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="runanywhere/SmolVLM-500M-Instruct-GGUF",
	filename="SmolVLM-500M-Instruct-Q8_0.gguf",
)

llm.create_chat_completion(
	messages = [
		{
			"role": "user",
			"content": [
				{
					"type": "text",
					"text": "Describe this image in one sentence."
				},
				{
					"type": "image_url",
					"image_url": {
						"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
					}
				}
			]
		}
	]
)

Notebooks
Google Colab
Kaggle
Local Apps

llama.cpp

How to use runanywhere/SmolVLM-500M-Instruct-GGUF with llama.cpp:

Install from brew

brew install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf runanywhere/SmolVLM-500M-Instruct-GGUF:Q8_0
# Run inference directly in the terminal:
llama-cli -hf runanywhere/SmolVLM-500M-Instruct-GGUF:Q8_0

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf runanywhere/SmolVLM-500M-Instruct-GGUF:Q8_0
# Run inference directly in the terminal:
llama-cli -hf runanywhere/SmolVLM-500M-Instruct-GGUF:Q8_0

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 runanywhere/SmolVLM-500M-Instruct-GGUF:Q8_0
# Run inference directly in the terminal:
./llama-cli -hf runanywhere/SmolVLM-500M-Instruct-GGUF:Q8_0

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 runanywhere/SmolVLM-500M-Instruct-GGUF:Q8_0
# Run inference directly in the terminal:
./build/bin/llama-cli -hf runanywhere/SmolVLM-500M-Instruct-GGUF:Q8_0

Use Docker

docker model run hf.co/runanywhere/SmolVLM-500M-Instruct-GGUF:Q8_0

LM Studio
Jan

vLLM

How to use runanywhere/SmolVLM-500M-Instruct-GGUF with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "runanywhere/SmolVLM-500M-Instruct-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": "runanywhere/SmolVLM-500M-Instruct-GGUF",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Use Docker

docker model run hf.co/runanywhere/SmolVLM-500M-Instruct-GGUF:Q8_0

Ollama
How to use runanywhere/SmolVLM-500M-Instruct-GGUF with Ollama:
```
ollama run hf.co/runanywhere/SmolVLM-500M-Instruct-GGUF:Q8_0
```

Unsloth Studio new

How to use runanywhere/SmolVLM-500M-Instruct-GGUF with Unsloth Studio:

Install Unsloth Studio (macOS, Linux, WSL)

curl -fsSL https://unsloth.ai/install.sh | sh
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for runanywhere/SmolVLM-500M-Instruct-GGUF to start chatting

Install Unsloth Studio (Windows)

irm https://unsloth.ai/install.ps1 | iex
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for runanywhere/SmolVLM-500M-Instruct-GGUF to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for runanywhere/SmolVLM-500M-Instruct-GGUF to start chatting

Docker Model Runner
How to use runanywhere/SmolVLM-500M-Instruct-GGUF with Docker Model Runner:
```
docker model run hf.co/runanywhere/SmolVLM-500M-Instruct-GGUF:Q8_0
```

Lemonade

How to use runanywhere/SmolVLM-500M-Instruct-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull runanywhere/SmolVLM-500M-Instruct-GGUF:Q8_0

Run and chat with the model

lemonade run user.SmolVLM-500M-Instruct-GGUF-Q8_0

List all available models

lemonade list

SmolVLM 500M Instruct — GGUF

SmolVLM-500M-Instruct quantized to GGUF format for llama.cpp, packaged for use with the RunAnywhere SDK.

Files:

SmolVLM-500M-Instruct-Q8_0.gguf — Language model (~416 MB)
mmproj-SmolVLM-500M-Instruct-f16.gguf — Vision encoder (~190 MB)

Usage with RunAnywhere SDK

Swift (iOS / macOS)

import RunAnywhere

RunAnywhere.registerModel(
    id: "smolvlm-500m-instruct-q8_0",
    name: "SmolVLM 500M Instruct Q8_0",
    repo: "runanywhere/SmolVLM-500M-Instruct-GGUF",
    files: ["SmolVLM-500M-Instruct-Q8_0.gguf", "mmproj-SmolVLM-500M-Instruct-f16.gguf"],
    framework: .llamaCpp,
    modality: .multimodal,
    memoryRequirement: 600_000_000
)

// VLM inference with image
let result = try await RunAnywhere.generateVLM(
    prompt: "Describe what you see in this image.",
    image: imageData,
    modelId: "smolvlm-500m-instruct-q8_0"
)

Kotlin (Android / JVM)

import com.runanywhere.sdk.RunAnywhere
import com.runanywhere.sdk.models.*

RunAnywhere.registerModel(
    id = "smolvlm-500m-instruct-q8_0",
    name = "SmolVLM 500M Instruct Q8_0",
    repo = "runanywhere/SmolVLM-500M-Instruct-GGUF",
    files = listOf("SmolVLM-500M-Instruct-Q8_0.gguf", "mmproj-SmolVLM-500M-Instruct-f16.gguf"),
    framework = InferenceFramework.LLAMA_CPP,
    modality = ModelCategory.MULTIMODAL,
    memoryRequirement = 600_000_000L
)

val result = RunAnywhere.generateVLM(
    prompt = "Describe what you see.",
    image = imageData,
    modelId = "smolvlm-500m-instruct-q8_0"
)

Web (TypeScript)

import { RunAnywhere, LLMFramework, ModelCategory } from '@anthropic/runanywhere-web';

RunAnywhere.registerModels([{
  id: 'smolvlm-500m-instruct-q8_0',
  name: 'SmolVLM 500M Instruct Q8_0',
  repo: 'runanywhere/SmolVLM-500M-Instruct-GGUF',
  files: ['SmolVLM-500M-Instruct-Q8_0.gguf', 'mmproj-SmolVLM-500M-Instruct-f16.gguf'],
  framework: LLMFramework.LlamaCpp,
  modality: ModelCategory.Multimodal,
  memoryRequirement: 600_000_000,
}]);

await RunAnywhere.downloadModel('smolvlm-500m-instruct-q8_0');
await RunAnywhere.loadModel('smolvlm-500m-instruct-q8_0');

const result = await RunAnywhere.generateVLM('Describe what you see.', imageData, 'smolvlm-500m-instruct-q8_0');

React Native (TypeScript)

import { RunAnywhere } from 'runanywhere-react-native';

RunAnywhere.registerModel({
  id: 'smolvlm-500m-instruct-q8_0',
  name: 'SmolVLM 500M Instruct Q8_0',
  repo: 'runanywhere/SmolVLM-500M-Instruct-GGUF',
  files: ['SmolVLM-500M-Instruct-Q8_0.gguf', 'mmproj-SmolVLM-500M-Instruct-f16.gguf'],
  framework: 'llamaCpp',
  modality: 'multimodal',
  memoryRequirement: 600_000_000,
});

const result = await RunAnywhere.generateVLM('Describe what you see.', imageData, 'smolvlm-500m-instruct-q8_0');

Flutter (Dart)

import 'package:runanywhere_flutter/runanywhere_flutter.dart';

RunAnywhere.registerModel(
  id: 'smolvlm-500m-instruct-q8_0',
  name: 'SmolVLM 500M Instruct Q8_0',
  repo: 'runanywhere/SmolVLM-500M-Instruct-GGUF',
  files: ['SmolVLM-500M-Instruct-Q8_0.gguf', 'mmproj-SmolVLM-500M-Instruct-f16.gguf'],
  framework: InferenceFramework.llamaCpp,
  modality: ModelCategory.multimodal,
  memoryRequirement: 600000000,
);

final result = await RunAnywhere.generateVLM('Describe what you see.', imageData, 'smolvlm-500m-instruct-q8_0');

Model Details

Property	Value
Base Model	SmolVLM-500M-Instruct
Parameters	500M
Quantization	Q8_0
Runtime	llama.cpp (with multimodal/mtmd)
Vision Encoder	SigLIP (F16)

Attribution

Original model by HuggingFace. GGUF conversion by ggml-org.

Downloads last month: 97

GGUF

Model size

0.4B params

Architecture

llama

Hardware compatibility

8-bit