Instructions to use tensorblock/granite-3.2-8b-instruct-preview-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use tensorblock/granite-3.2-8b-instruct-preview-GGUF with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="tensorblock/granite-3.2-8b-instruct-preview-GGUF")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoModel
model = AutoModel.from_pretrained("tensorblock/granite-3.2-8b-instruct-preview-GGUF", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

llama.cpp

How to use tensorblock/granite-3.2-8b-instruct-preview-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 tensorblock/granite-3.2-8b-instruct-preview-GGUF:Q2_K
# Run inference directly in the terminal:
llama cli -hf tensorblock/granite-3.2-8b-instruct-preview-GGUF:Q2_K

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf tensorblock/granite-3.2-8b-instruct-preview-GGUF:Q2_K
# Run inference directly in the terminal:
llama cli -hf tensorblock/granite-3.2-8b-instruct-preview-GGUF:Q2_K

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 tensorblock/granite-3.2-8b-instruct-preview-GGUF:Q2_K
# Run inference directly in the terminal:
./llama-cli -hf tensorblock/granite-3.2-8b-instruct-preview-GGUF:Q2_K

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 tensorblock/granite-3.2-8b-instruct-preview-GGUF:Q2_K
# Run inference directly in the terminal:
./build/bin/llama-cli -hf tensorblock/granite-3.2-8b-instruct-preview-GGUF:Q2_K

Use Docker

docker model run hf.co/tensorblock/granite-3.2-8b-instruct-preview-GGUF:Q2_K

LM Studio
Jan

vLLM

How to use tensorblock/granite-3.2-8b-instruct-preview-GGUF with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "tensorblock/granite-3.2-8b-instruct-preview-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": "tensorblock/granite-3.2-8b-instruct-preview-GGUF",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/tensorblock/granite-3.2-8b-instruct-preview-GGUF:Q2_K

SGLang

How to use tensorblock/granite-3.2-8b-instruct-preview-GGUF with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "tensorblock/granite-3.2-8b-instruct-preview-GGUF" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "tensorblock/granite-3.2-8b-instruct-preview-GGUF",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "tensorblock/granite-3.2-8b-instruct-preview-GGUF" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "tensorblock/granite-3.2-8b-instruct-preview-GGUF",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Ollama
How to use tensorblock/granite-3.2-8b-instruct-preview-GGUF with Ollama:
```
ollama run hf.co/tensorblock/granite-3.2-8b-instruct-preview-GGUF:Q2_K
```
Unsloth Desktop

How to use tensorblock/granite-3.2-8b-instruct-preview-GGUF with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf tensorblock/granite-3.2-8b-instruct-preview-GGUF:Q2_K

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": "tensorblock/granite-3.2-8b-instruct-preview-GGUF:Q2_K"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use tensorblock/granite-3.2-8b-instruct-preview-GGUF with Docker Model Runner:
```
docker model run hf.co/tensorblock/granite-3.2-8b-instruct-preview-GGUF:Q2_K
```

Lemonade

How to use tensorblock/granite-3.2-8b-instruct-preview-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull tensorblock/granite-3.2-8b-instruct-preview-GGUF:Q2_K

Run and chat with the model

lemonade run user.granite-3.2-8b-instruct-preview-GGUF-Q2_K

List all available models

lemonade list

Hermes Agent

How to use tensorblock/granite-3.2-8b-instruct-preview-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 tensorblock/granite-3.2-8b-instruct-preview-GGUF:Q2_K

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 tensorblock/granite-3.2-8b-instruct-preview-GGUF:Q2_K

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use tensorblock/granite-3.2-8b-instruct-preview-GGUF with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf tensorblock/granite-3.2-8b-instruct-preview-GGUF:Q2_K

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 "tensorblock/granite-3.2-8b-instruct-preview-GGUF:Q2_K" \
  --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"

A newer version of this model is available: ibm-granite/granite-3.3-8b-instruct

ibm-granite/granite-3.2-8b-instruct-preview - GGUF

This repo contains GGUF format model files for ibm-granite/granite-3.2-8b-instruct-preview.

The files were quantized using machines provided by TensorBlock, and they are compatible with llama.cpp as of commit b4823.

Our projects

Forge

An OpenAI-compatible multi-provider routing layer.
🚀 Try it now! 🚀
Awesome MCP Servers	TensorBlock Studio

A comprehensive collection of Model Context Protocol (MCP) servers.	A lightweight, open, and extensible multi-LLM interaction studio.
👀 See what we built 👀	👀 See what we built 👀

## Prompt template

<|start_of_role|>system<|end_of_role|>{system_prompt}<|end_of_text|>
<|start_of_role|>user<|end_of_role|>{prompt}<|end_of_text|>
<|start_of_role|>assistant<|end_of_role|>

Model file specification

Filename	Quant type	File Size	Description
granite-3.2-8b-instruct-preview-Q2_K.gguf	Q2_K	3.104 GB	smallest, significant quality loss - not recommended for most purposes
granite-3.2-8b-instruct-preview-Q3_K_S.gguf	Q3_K_S	3.592 GB	very small, high quality loss
granite-3.2-8b-instruct-preview-Q3_K_M.gguf	Q3_K_M	3.997 GB	very small, high quality loss
granite-3.2-8b-instruct-preview-Q3_K_L.gguf	Q3_K_L	4.349 GB	small, substantial quality loss
granite-3.2-8b-instruct-preview-Q4_0.gguf	Q4_0	4.651 GB	legacy; small, very high quality loss - prefer using Q3_K_M
granite-3.2-8b-instruct-preview-Q4_K_S.gguf	Q4_K_S	4.686 GB	small, greater quality loss
granite-3.2-8b-instruct-preview-Q4_K_M.gguf	Q4_K_M	4.943 GB	medium, balanced quality - recommended
granite-3.2-8b-instruct-preview-Q5_0.gguf	Q5_0	5.647 GB	legacy; medium, balanced quality - prefer using Q4_K_M
granite-3.2-8b-instruct-preview-Q5_K_S.gguf	Q5_K_S	5.647 GB	large, low quality loss - recommended
granite-3.2-8b-instruct-preview-Q5_K_M.gguf	Q5_K_M	5.797 GB	large, very low quality loss - recommended
granite-3.2-8b-instruct-preview-Q6_K.gguf	Q6_K	6.705 GB	very large, extremely low quality loss
granite-3.2-8b-instruct-preview-Q8_0.gguf	Q8_0	8.684 GB	very large, extremely low quality loss - not recommended

Downloading instruction

Command line

Firstly, install Huggingface Client

pip install -U "huggingface_hub[cli]"

Then, downoad the individual model file the a local directory

huggingface-cli download tensorblock/granite-3.2-8b-instruct-preview-GGUF --include "granite-3.2-8b-instruct-preview-Q2_K.gguf" --local-dir MY_LOCAL_DIR

If you wanna download multiple model files with a pattern (e.g., *Q4_K*gguf), you can try:

huggingface-cli download tensorblock/granite-3.2-8b-instruct-preview-GGUF --local-dir MY_LOCAL_DIR --local-dir-use-symlinks False --include='*Q4_K*gguf'

Downloads last month: 113

GGUF

Model size

8B params

Architecture

granite

Hardware compatibility

2-bit

3-bit

Model tree for tensorblock/granite-3.2-8b-instruct-preview-GGUF

Base model

ibm-granite/granite-3.1-8b-base

Finetuned

ibm-granite/granite-3.1-8b-instruct

Finetuned

ibm-granite/granite-3.2-8b-instruct-preview

Quantized

(13)

this model