Instructions to use AI-Sweden-Models/gpt-sw3-1.3b-instruct with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use AI-Sweden-Models/gpt-sw3-1.3b-instruct with Transformers:

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

pipe = pipeline("text-generation", model="AI-Sweden-Models/gpt-sw3-1.3b-instruct")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("AI-Sweden-Models/gpt-sw3-1.3b-instruct")
model = AutoModelForCausalLM.from_pretrained("AI-Sweden-Models/gpt-sw3-1.3b-instruct", device_map="auto")
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=256)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use AI-Sweden-Models/gpt-sw3-1.3b-instruct with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "AI-Sweden-Models/gpt-sw3-1.3b-instruct"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "AI-Sweden-Models/gpt-sw3-1.3b-instruct",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/AI-Sweden-Models/gpt-sw3-1.3b-instruct

SGLang

How to use AI-Sweden-Models/gpt-sw3-1.3b-instruct 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 "AI-Sweden-Models/gpt-sw3-1.3b-instruct" \
    --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": "AI-Sweden-Models/gpt-sw3-1.3b-instruct",
		"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 "AI-Sweden-Models/gpt-sw3-1.3b-instruct" \
        --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": "AI-Sweden-Models/gpt-sw3-1.3b-instruct",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use AI-Sweden-Models/gpt-sw3-1.3b-instruct with Docker Model Runner:
```
docker model run hf.co/AI-Sweden-Models/gpt-sw3-1.3b-instruct
```

translation

#10

by exoplanet - opened Oct 19, 2025

Discussion

exoplanet

Oct 19, 2025

Hi team,
Would you recommend using this model for translating [ sv da nb ] to/from en ?
Cheers.

timpal0l

AI Sweden Model Hub org Oct 23, 2025

Yeah, try it, even the 6.7b models if you can.

Also check out: https://huggingface.co/AI-Sweden-Models/gpt-sw3-6.7b-v2-translator

exoplanet

Oct 24, 2025

•

edited Oct 25, 2025

Oh, if only my target environment had more memory, that translator fine-tuned model would be awesome, however up to 4B params would fit the bill. As I'm not native a speaker of sv / da / nb, I cannot tell whether the translations are good or somewhat hallucinated. Therefore, should I opt for a smaller translation fine-tuned models instead?

Kenopet

Nov 17, 2025

•

edited Nov 17, 2025

I have tried sv-en to/from using AI-Sweden (both GPT 1.3b and 6.7b v2 translator), X-ALMA and Helsinki-NLP. This model works good enough for translation but the 6.7b v2 Translator is way better in niche situations, but still NOT good on Industry specific language (None of the models are). X-LLama is similar to AI-Sweden/GPT sw3 6.7b v2 translator might be somewhat better but that expected from such a larger model.

The super lightweight one way translators from Helsinki-NLP are better than GPT 1.3b but worse compared to the bigger models. So if your limited in compute space (not storage due to needing 6 models from the Helsinki-NLP for to/from on 3 languages) then its a good choice. https://huggingface.co/Helsinki-NLP

exoplanet

Nov 21, 2025

Thanks @Kenopet for these great insights!

exoplanet changed discussion status to closed Nov 21, 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