Instructions to use LumiOpen/Llama-Poro-2-8B-Instruct with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use LumiOpen/Llama-Poro-2-8B-Instruct with Transformers:

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

pipe = pipeline("text-generation", model="LumiOpen/Llama-Poro-2-8B-Instruct")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("LumiOpen/Llama-Poro-2-8B-Instruct")
model = AutoModelForCausalLM.from_pretrained("LumiOpen/Llama-Poro-2-8B-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=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Inference
Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use LumiOpen/Llama-Poro-2-8B-Instruct with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "LumiOpen/Llama-Poro-2-8B-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": "LumiOpen/Llama-Poro-2-8B-Instruct",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/LumiOpen/Llama-Poro-2-8B-Instruct

SGLang

How to use LumiOpen/Llama-Poro-2-8B-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 "LumiOpen/Llama-Poro-2-8B-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": "LumiOpen/Llama-Poro-2-8B-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 "LumiOpen/Llama-Poro-2-8B-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": "LumiOpen/Llama-Poro-2-8B-Instruct",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use LumiOpen/Llama-Poro-2-8B-Instruct with Docker Model Runner:
```
docker model run hf.co/LumiOpen/Llama-Poro-2-8B-Instruct
```

Suomen kielen osaaminen

by ArttuPakarinen - opened Feb 21

Discussion

ArttuPakarinen

Feb 21

Osaa suomen kieltä ihan hyvin, suurimmaksi osaksi.000

Mutta!

Tyypillisiä AI vastauksia jota tämä tuottaa:

Minä olen täällä auttamaan. "I'm here to help.."
Mitä sinä kuuluu? "How are you doing?"

Onko nää jo treenausvaiheessa tulleita käännöskukkasia? Tuntuu juuri näitä, erityisesti 1. toistavan niin useasti.

mpasila

Feb 23

Jos katsot tän modellin datasetin readmetä niin siinä ne mainitsi että ne käänsi suurimman osan datasta Llama 3.3 70B ja Poro 34B modelleilla englannista suomeksi.

LumiOpen/poro2-instruction-collection

# Finnish data 

For the Finnish portion, we translated prompts from the Tulu3 SFT Mixture into Finnish. We used Llama-3.3-70B-Instruct to generate multiple responses to the translated prompts and used the same model to select the best response. 
We supplemented this data with the top Finnish conversations from Open Assistant 2 and Avoin Avustaja.

### Prompt selection

We deduplicated the prompts in the Tulu3 dataset, excluded prompts that have a non-commercial license, and excluded prompts that are not in English.

### Prompt translation

We translated the prompts with Poro-34B using few-shot prompting. Following our previous work, we experimented with different prompt formats and number of examples.

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