Instructions to use cortexso/hermes3 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use cortexso/hermes3 with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="cortexso/hermes3",
	filename="hermes-3-llama-3.2-3b-q2_k.gguf",
)

llm.create_chat_completion(
	messages = [
		{
			"role": "user",
			"content": "What is the capital of France?"
		}
	]
)

Notebooks
Google Colab
Kaggle
Local Apps

llama.cpp

How to use cortexso/hermes3 with llama.cpp:

Install from brew

brew install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf cortexso/hermes3:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf cortexso/hermes3:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf cortexso/hermes3:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf cortexso/hermes3: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 cortexso/hermes3:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf cortexso/hermes3: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 cortexso/hermes3:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf cortexso/hermes3:Q4_K_M

Use Docker

docker model run hf.co/cortexso/hermes3:Q4_K_M

LM Studio
Jan

vLLM

How to use cortexso/hermes3 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "cortexso/hermes3"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "cortexso/hermes3",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/cortexso/hermes3:Q4_K_M

Ollama
How to use cortexso/hermes3 with Ollama:
```
ollama run hf.co/cortexso/hermes3:Q4_K_M
```

Unsloth Studio new

How to use cortexso/hermes3 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 cortexso/hermes3 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 cortexso/hermes3 to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for cortexso/hermes3 to start chatting

Docker Model Runner
How to use cortexso/hermes3 with Docker Model Runner:
```
docker model run hf.co/cortexso/hermes3:Q4_K_M
```

Lemonade

How to use cortexso/hermes3 with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull cortexso/hermes3:Q4_K_M

Run and chat with the model

lemonade run user.hermes3-Q4_K_M

List all available models

lemonade list

Overview

Nous Research developed and released the Hermes 3, a state-of-the-art instruction-tuned language model built on Llama-3.2-3B. This 3-billion parameter model is a fine-tuned version of Llama-3.2 and represents a leap forward in reasoning, multi-turn conversation, and structured outputs. It incorporates advanced role-playing capabilities, reliable function calling, and improved coherence over long contexts, making it a versatile assistant for various applications.

Hermes 3 was trained with high-quality data, leveraging fine-tuning techniques on H100 GPUs via LambdaLabs GPU Cloud. The model excels in both general-purpose and specialized tasks, including code generation, reasoning, and advanced conversational abilities. With support for ChatML prompt formatting, Hermes 3 ensures compatibility with OpenAI endpoints and facilitates structured, steerable interactions for end-users.

Variants

No	Variant	Cortex CLI command
1	Hermes3-3b	`cortex run hermes3:3b`

Use it with Jan (UI)

Install Jan using Quickstart
Use in Jan model Hub:
```
cortexso/hermes3
```

Use it with Cortex (CLI)

Install Cortex using Quickstart
Run the model with command:
```
cortex run hermes3
```

Credits

Author: Nous Research
Converter: Homebrew
Original License: License
Papers: Hermes 3 Technical Report

Downloads last month: 132

GGUF

Model size

3B params

Architecture

llama

Hardware compatibility

2-bit

3-bit

4-bit

5-bit

6-bit

8-bit

Paper for cortexso/hermes3

Hermes 3 Technical Report

Paper • 2408.11857 • Published Aug 15, 2024 • 58