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

Libraries

How to use meta-llama/Meta-Llama-3-8B-Instruct with Transformers:

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

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

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("meta-llama/Meta-Llama-3-8B-Instruct")
model = AutoModelForCausalLM.from_pretrained("meta-llama/Meta-Llama-3-8B-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 meta-llama/Meta-Llama-3-8B-Instruct with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "meta-llama/Meta-Llama-3-8B-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": "meta-llama/Meta-Llama-3-8B-Instruct",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/meta-llama/Meta-Llama-3-8B-Instruct

SGLang

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

Docker Model Runner
How to use meta-llama/Meta-Llama-3-8B-Instruct with Docker Model Runner:
```
docker model run hf.co/meta-llama/Meta-Llama-3-8B-Instruct
```

Max output tokens?

#12

by stri8ted - opened Apr 18, 2024

Discussion

stri8ted

Apr 18, 2024

I understand the input context length is 8k, but what about the output?

npotts

Apr 19, 2024

•

edited Apr 19, 2024

The output takes the same space as the input. The model completes the 8k token space with the response. You could slide the contxt window and get more output, but then you're losing context at the beginning.

Example, if your input is 100 tokens, you have ~7900 tokens for completion. But if your input is 7900 tokens, you have ~100 tokens left for response until your input starts getting trimmed out. The model can only pay attention to 8k tokens at most.

YalunHu

May 9, 2024

I understand the input context length is 8k, but what about the output?

May I ask where I can check this "8k context length" configuration for the llama3 model? Thanks!

npotts

May 9, 2024

I understand the input context length is 8k, but what about the output?

May I ask where I can check this "8k context length" configuration for the llama3 model? Thanks!

https://huggingface.co/meta-llama/Meta-Llama-3-8B-Instruct/blob/main/config.json

look at max_position_embeddings

averyyu99

Dec 13, 2024

The output takes the same space as the input. The model completes the 8k token space with the response. You could slide the contxt window and get more output, but then you're losing context at the beginning.

Example, if your input is 100 tokens, you have ~7900 tokens for completion. But if your input is 7900 tokens, you have ~100 tokens left for response until your input starts getting trimmed out. The model can only pay attention to 8k tokens at most.

Thanks for you detailed explanation! However, I am wondering if the prompt length was count towards the 8000 tokens limit? As I see "system

Cutting Knowledge Date: December 2023
Today Date: 26 Jul 2024 <Prompt...>" in my output

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