Instructions to use abacusai/Llama-3-Smaug-8B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use abacusai/Llama-3-Smaug-8B with Transformers:

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

pipe = pipeline("text-generation", model="abacusai/Llama-3-Smaug-8B")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("abacusai/Llama-3-Smaug-8B")
model = AutoModelForCausalLM.from_pretrained("abacusai/Llama-3-Smaug-8B")
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 abacusai/Llama-3-Smaug-8B with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "abacusai/Llama-3-Smaug-8B"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "abacusai/Llama-3-Smaug-8B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/abacusai/Llama-3-Smaug-8B

SGLang

How to use abacusai/Llama-3-Smaug-8B 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 "abacusai/Llama-3-Smaug-8B" \
    --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": "abacusai/Llama-3-Smaug-8B",
		"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 "abacusai/Llama-3-Smaug-8B" \
        --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": "abacusai/Llama-3-Smaug-8B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use abacusai/Llama-3-Smaug-8B with Docker Model Runner:
```
docker model run hf.co/abacusai/Llama-3-Smaug-8B
```

Any news about the updated model with the latest recipe?

by ManniX-ITA - opened Jun 17, 2024

Discussion

ManniX-ITA

Jun 17, 2024

@ArkaAbacus
Any news about when will be eventually released?
I'd like to add some more quants in the ollama library but as said I'd like to avoid it doing twice.

The latest 70b with 32k context seems better than the model without.
Would be nice if also the 8B could come with the 32k context exetension.

Thanks!

ArkaAbacus

Jun 26, 2024

•

edited Jun 26, 2024

Hi there, sorry for the slow response on this. 32k context 8B is not on our immediate roadmap. I could try to squeeze this in potentially in the next week or so, but I wouldn't necessarily count on it happening for sure.

ManniX-ITA

Jun 26, 2024

No problem, I have uploaded it already. When it comes out, I'll be glad to upload the new one!

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