Instructions to use ATH-MaaS/Marco-o1 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use ATH-MaaS/Marco-o1 with Transformers:

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

pipe = pipeline("text-generation", model="ATH-MaaS/Marco-o1")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("ATH-MaaS/Marco-o1")
model = AutoModelForCausalLM.from_pretrained("ATH-MaaS/Marco-o1")
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
Local Apps Settings

vLLM

How to use ATH-MaaS/Marco-o1 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "ATH-MaaS/Marco-o1"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "ATH-MaaS/Marco-o1",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/ATH-MaaS/Marco-o1

SGLang

How to use ATH-MaaS/Marco-o1 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 "ATH-MaaS/Marco-o1" \
    --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": "ATH-MaaS/Marco-o1",
		"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 "ATH-MaaS/Marco-o1" \
        --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": "ATH-MaaS/Marco-o1",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use ATH-MaaS/Marco-o1 with Docker Model Runner:
```
docker model run hf.co/ATH-MaaS/Marco-o1
```

A detailed tutorial for learning o1-like model

#17

by jitianbo - opened Nov 25, 2024

Discussion

jitianbo

Nov 25, 2024

Hi guys, I am new to o1, and I found this research is very helpful for learning how to play with an o1-like model! In addition, I found that Open-O1(https://github.com/Open-Source-O1/Open-O1) and O1-Journey (https://github.com/GAIR-NLP/O1-Journey) are useful as well. BTW, is there any way to contribute to this project?

longyuewang

ATH-MaaS org Nov 25, 2024

Hi, thx for your interests.
The projects you mentioned are concurrent with our research work (towards o1-like models). Maybe you find that there are similarities, such as in CoT SFT training, but we differ by focusing more on open-ended solutions. We've also observed some interesting phenomena in multilingual applications. I believe all of these efforts are still in the early stages.
We welcome more contributors!

Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.

Tap or paste here to upload images

· Sign up or log in to comment