Instructions to use Finnish-NLP/Ahma-7B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Finnish-NLP/Ahma-7B with Transformers:

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

pipe = pipeline("text-generation", model="Finnish-NLP/Ahma-7B")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("Finnish-NLP/Ahma-7B")
model = AutoModelForCausalLM.from_pretrained("Finnish-NLP/Ahma-7B")
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]:]))

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use Finnish-NLP/Ahma-7B with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Finnish-NLP/Ahma-7B"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Finnish-NLP/Ahma-7B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/Finnish-NLP/Ahma-7B

SGLang

How to use Finnish-NLP/Ahma-7B 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 "Finnish-NLP/Ahma-7B" \
    --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": "Finnish-NLP/Ahma-7B",
		"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 "Finnish-NLP/Ahma-7B" \
        --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": "Finnish-NLP/Ahma-7B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use Finnish-NLP/Ahma-7B with Docker Model Runner:
```
docker model run hf.co/Finnish-NLP/Ahma-7B
```

Improve model card: Add `transformers` library, link paper, include abstract

by nielsr HF Staff - opened Jul 1, 2025

base: refs/heads/main

←

from: refs/pr/1

Discussion Files changed

+92

-88

nielsr

Jul 1, 2025

This PR significantly enhances the model card for Ahma-7B by:

Adding library_name: transformers to the metadata: This ensures the Hugging Face Hub correctly recognizes the model's compatible library, enabling the "how to use" button and providing relevant code snippets for users.
Linking to the associated research paper: The model card now explicitly references "Scaling Data-Constrained Language Models", which describes the training strategy and research behind the Ahma model. This link is added to the introductory section and updated in the "2-stage pretraining" section for clarity.
Including the paper abstract: A dedicated "Paper Abstract" section has been added to provide users with immediate context about the research, its motivations, and key findings directly within the model card.
Removing inference: false from metadata: This tag was contradictory, as the model card provides clear inference usage examples. Removing it clarifies that the model is indeed ready for direct inference.

Improve model card: Add `transformers` library, link paper, include abstract6059cd3e

RASMUS changed pull request status to merged Jul 3, 2025

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