Instructions to use katanemo/Arch-Router-1.5B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use katanemo/Arch-Router-1.5B with Transformers:

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

pipe = pipeline("text-generation", model="katanemo/Arch-Router-1.5B")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("katanemo/Arch-Router-1.5B")
model = AutoModelForCausalLM.from_pretrained("katanemo/Arch-Router-1.5B")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Inference
HuggingChat
Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use katanemo/Arch-Router-1.5B with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "katanemo/Arch-Router-1.5B"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "katanemo/Arch-Router-1.5B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/katanemo/Arch-Router-1.5B

SGLang

How to use katanemo/Arch-Router-1.5B 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 "katanemo/Arch-Router-1.5B" \
    --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": "katanemo/Arch-Router-1.5B",
		"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 "katanemo/Arch-Router-1.5B" \
        --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": "katanemo/Arch-Router-1.5B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use katanemo/Arch-Router-1.5B with Docker Model Runner:
```
docker model run hf.co/katanemo/Arch-Router-1.5B
```

parachas commited on Jun 17, 2025

Commit

51fb37c

verified ·

1 Parent(s): 1bef53d

Update README.md

Browse files

Files changed (1) hide show

README.md +6 -3

README.md CHANGED Viewed

@@ -14,11 +14,14 @@ library_name: transformers
 # katanemo/Arch-Router-1.5B
 ## Overview
-With the rapid proliferation of large language models (LLM) — each optimized for different strengths, style, or latency/cost profile — routing has become an essential technique to operationalize the use of different LLMs.
-Arch-Router is a **preference-based routing model** designed to intelligently select the most appropriate large language model (LLM) for a given prompt by leveraging a structured **domain–action taxonomy**.
-This framework enables fine-grained control over model selection by aligning user-defined preferences with model capabilities across a wide range of tasks and subject areas - and matches practical definitions of performance in the real world to make routing decisions more transparent and adaptable.
 ### How It Works

 # katanemo/Arch-Router-1.5B
 ## Overview
+With the rapid proliferation of large language models (LLM)—each optimized for different strengths, style, or latency/cost profile—routing has become an essential technique to operationalize the use of different models.
+Existing work on LLM routing typically focuses on learning an optimal policy to route between a limited pool of models, where optimal is measured via well-defined performance benchmarks. This framework, however, is misaligned with real-world scenarios.
+Benchmark performance does not capture subjective evaluation and testing criteria in the real world.
+Arch-Router is a **preference-based routing model** designed to intelligently guide model selection by matching queries to user-defined domains (e.g., finance and healthcare) and action types (e.g., code generation, image editing, etc.).
+Experiments on conversational datasets demonstrate that our approach achieves state-of-the-art (SOTA) results in matching queries with human preferences, outperforming top proprietary routing systems.
+Our preference-aligned approach matches practical definitions of performance in the real world and makes routing decisions more transparent and adaptable.
 ### How It Works