Instructions to use raymsm/Llama-3.2-3B-Instruct-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- llama-cpp-python
How to use raymsm/Llama-3.2-3B-Instruct-GGUF with llama-cpp-python:
# !pip install llama-cpp-python from llama_cpp import Llama llm = Llama.from_pretrained( repo_id="raymsm/Llama-3.2-3B-Instruct-GGUF", filename="Llama-3.2-3B-Instruct.Q4_K_M-imat.gguf", )
llm.create_chat_completion( messages = "No input example has been defined for this model task." )
- Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- llama.cpp
How to use raymsm/Llama-3.2-3B-Instruct-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 raymsm/Llama-3.2-3B-Instruct-GGUF:Q4_K_M # Run inference directly in the terminal: llama cli -hf raymsm/Llama-3.2-3B-Instruct-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 raymsm/Llama-3.2-3B-Instruct-GGUF:Q4_K_M # Run inference directly in the terminal: llama cli -hf raymsm/Llama-3.2-3B-Instruct-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 raymsm/Llama-3.2-3B-Instruct-GGUF:Q4_K_M # Run inference directly in the terminal: ./llama-cli -hf raymsm/Llama-3.2-3B-Instruct-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 raymsm/Llama-3.2-3B-Instruct-GGUF:Q4_K_M # Run inference directly in the terminal: ./build/bin/llama-cli -hf raymsm/Llama-3.2-3B-Instruct-GGUF:Q4_K_M
Use Docker
docker model run hf.co/raymsm/Llama-3.2-3B-Instruct-GGUF:Q4_K_M
- LM Studio
- Jan
- Ollama
How to use raymsm/Llama-3.2-3B-Instruct-GGUF with Ollama:
ollama run hf.co/raymsm/Llama-3.2-3B-Instruct-GGUF:Q4_K_M
- Unsloth Studio
How to use raymsm/Llama-3.2-3B-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 raymsm/Llama-3.2-3B-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 raymsm/Llama-3.2-3B-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 raymsm/Llama-3.2-3B-Instruct-GGUF to start chatting
- Pi
How to use raymsm/Llama-3.2-3B-Instruct-GGUF with Pi:
Start the llama.cpp server
# Install llama.cpp: brew install llama.cpp # Start a local OpenAI-compatible server: llama serve -hf raymsm/Llama-3.2-3B-Instruct-GGUF:Q4_K_M
Configure the model in Pi
# Install Pi: npm install -g @mariozechner/pi-coding-agent # Add to ~/.pi/agent/models.json: { "providers": { "llama-cpp": { "baseUrl": "http://localhost:8080/v1", "api": "openai-completions", "apiKey": "none", "models": [ { "id": "raymsm/Llama-3.2-3B-Instruct-GGUF:Q4_K_M" } ] } } }Run Pi
# Start Pi in your project directory: pi
- Hermes Agent new
How to use raymsm/Llama-3.2-3B-Instruct-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 raymsm/Llama-3.2-3B-Instruct-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 raymsm/Llama-3.2-3B-Instruct-GGUF:Q4_K_M
Run Hermes
hermes
- Atomic Chat new
- OpenClaw new
How to use raymsm/Llama-3.2-3B-Instruct-GGUF with OpenClaw:
Start the llama.cpp server
# Install llama.cpp: brew install llama.cpp # Start a local OpenAI-compatible server: llama serve -hf raymsm/Llama-3.2-3B-Instruct-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 "raymsm/Llama-3.2-3B-Instruct-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"
- Docker Model Runner
How to use raymsm/Llama-3.2-3B-Instruct-GGUF with Docker Model Runner:
docker model run hf.co/raymsm/Llama-3.2-3B-Instruct-GGUF:Q4_K_M
- Lemonade
How to use raymsm/Llama-3.2-3B-Instruct-GGUF with Lemonade:
Pull the model
# Download Lemonade from https://lemonade-server.ai/ lemonade pull raymsm/Llama-3.2-3B-Instruct-GGUF:Q4_K_M
Run and chat with the model
lemonade run user.Llama-3.2-3B-Instruct-GGUF-Q4_K_M
List all available models
lemonade list
YAML Metadata Warning:empty or missing yaml metadata in repo card
Check out the documentation for more information.
Llama-3.2-3B-Instruct GGUF
Quantized GGUF versions of meta-llama/Llama-3.2-3B-Instruct generated using llama.cpp on an Oracle Cloud VM.
Repositories & Quantization Details
This repository contains the following quantization formats:
| Quantization Type | File Size | Quantization Duration | Peak RAM Used | Description |
|---|---|---|---|---|
| Q4_K_M | 1.88 GB | 32.5s | 9194.1 MB | Medium-sized, good balance of size/perplexity. |
| Q5_K_M | 2.16 GB | 25.7s | 9193.8 MB | Medium-sized, good balance of size/perplexity. |
| Q6_K | 2.46 GB | 19.1s | 9194.1 MB | Medium-sized, good balance of size/perplexity. |
| Q8_0 | 3.19 GB | 9.8s | 9113.6 MB | Medium-sized, good balance of size/perplexity. |
Sanity Check Prompts & Responses
Here are the responses generated by the quantized models to verify coherent text generation.
Format: Q4_K_M
- Prompt: What is the capital of France? Answer in one word.
- Response:
What is the capital of France? Answer in one word. Paris. What is the capital of France? Answer in one word. Paris. What is the capital of France? Answer in one word. Paris. What is the capital of France? Answer in one word. Paris. What is the capital of France? Answer in one word. Paris. What - Prompt: Write a short four-line poem about coding in Python.
- Response:
Write a short four-line poem about coding in Python. Here is the list of words to include in the list of words to include in the the poem: - Python
- Indentation
- Loops
- Variables
Here is the the list of words to include in the list of words to include in the the poem:
- Indent
- Prompt: Explain why the sky is blue in one simple sentence.
- Response:
Explain why the sky is blue in one simple sentence. The sky appears blue because of a phenomenon called Rayleigh scattering, in which shorter wavelengths of light (like blue and violet) are scattered more than longer wavelengths (like red and orange) by the tiny molecules of gases in the Earth's atmosphere. This scattering effect gives the sky its blue color.
Note: This
### Format: Q5_K_M
- **Prompt**: *What is the capital of France? Answer in one word.*
- **Response**:
```text
What is the capital of France? Answer in one word. Paris.
What is the capital of France? Answer in one word. Paris.
The answer is Paris. The question was answered correctly in both instances. The response should be a single word, "Paris". The response should be a single word, "Paris". The answer is Paris. The question was answered correctly in both
- Prompt: Write a short four-line poem about coding in Python.
- Response:
Write a short four-line poem about coding in Python. Here is the list of words to include in the list of words to include in the `words` list: - Python
- coding
- syntax
- indentation
- loops
- functions
- variables
- data
- type
- error
- debug
Here is the list of words to include
- **Prompt**: *Explain why the sky is blue in one simple sentence.*
- **Response**:
```text
Explain why the sky is blue in one simple sentence.
The sky appears blue because of a phenomenon called Rayleigh scattering, in which shorter wavelengths of light (like blue and violet) are scattered more than longer wavelengths (like red and orange) by the tiny molecules of gases in the Earth's atmosphere. This scattering effect gives the sky its blue color. [end of text]
Format: Q6_K
- Prompt: What is the capital of France? Answer in one word.
- Response:
What is the capital of France? Answer in one word. Paris. [end of text] - Prompt: Write a short four-line poem about coding in Python.
- Response:
Write a short four-line poem about coding in Python. Here is the list of words you can use: - Indent
- Indentation
- Syntax
- Loop
- Indentation error
- Indentation errors
- Indentation errors occur
- Indentation errors occur when
- Indentation errors occur when you
- Ind
- Prompt: Explain why the sky is blue in one simple sentence.
- Response:
Explain why the sky is blue in one simple sentence. The sky appears blue because of a phenomenon called Rayleigh scattering, in which shorter wavelengths of light (like blue and violet) are scattered more than longer wavelengths (like red and orange) by the tiny molecules of gases in the Earth's atmosphere.
(Note: This is a simplified explanation, but it's a good
### Format: Q8_0
- **Prompt**: *What is the capital of France? Answer in one word.*
- **Response**:
```text
What is the capital of France? Answer in one word.etworketwork comun comun comun comun comun comun comun comun comun comun comun comun comunciciciciciciciciciciciciciciciciciciciciciciacketacketacketacketacketacketacketacketacketacketacketacketacketacketacketacketacketacketacketacketacketacketacketacketacketacketacket
- Prompt: Write a short four-line poem about coding in Python.
- Response:
Write a short four-line poem about coding in Python.alararcetworketworketworketworketworketworketworkottiottiottiottiottiottiottiottiottiottiottiottiottiottiottiottiottiottiottiottiottiottiottiinicinicinicadcadcadcadcadcadcadcadcadcadcadcadcadcadcadcadcadcadcadcadcadcadcadc zap zap zap zap zap zap - Prompt: Explain why the sky is blue in one simple sentence.
- Response:
Explain why the sky is blue in one simple sentence.izonizonizonizonizonizonizonizonizonizonizonizon ordinarily ordinarily stat stat stat stat stat stat stat stat stat stat stat stat stat stat stat stat stat spinning spun spun spinning spinning spinning spinning spinning spinning spinning spinning spinning spinning spinning spinning spinning spinning spinning spinning spinning spinning spinning spinning spinning performer performer performer performer performer performer performer performer performer
Pipeline Logs
- Base GGUF Conversion Peak RAM: 3010.7 MB, Duration: 39.3s.
- Logged at:
Fri Jun 19 17:40:13 2026
Imatrix-Calibrated Quantization (Domain: Materials Science/Chemistry)
We calculated an importance matrix from cond-mat/chemistry abstracts scraped from arXiv (physics.chem-ph and cond-mat.mtrl-sci) to optimize quantization layers. Below is the perplexity comparison evaluated on a held-out slice of the same corpus (lower perplexity is better).
Perplexity Evaluation Results (50 chunks)
| Model Variant | Perplexity (PPL) | Description |
|---|---|---|
| Base (FP16/BF16) | 14.6012 | Unquantized baseline model reference. |
| Standard Q4_K_M | 14.9283 | Standard 4-bit k-quant without calibration. |
| Imatrix Q4_K_M | 14.7100 | Imatrix-calibrated 4-bit k-quant (Q4_K_M-imat.gguf). |
| Standard Q5_K_M | 14.6729 | Standard 5-bit k-quant without calibration. |
| Imatrix Q5_K_M | 14.7618 | Imatrix-calibrated 5-bit k-quant (Q5_K_M-imat.gguf). |
- Calibration Corpus Source: ArXiv abstracts (Materials Science & Chemistry)
- Calibration Corpus Size: ~5.24 MB (4,799 abstracts)
- Evaluation Corpus Size: ~1.33 MB (1,200 abstracts)
- Computed at:
Sat Jun 20 02:37:15 2026
- Downloads last month
- 60
4-bit
5-bit
6-bit
8-bit