Instructions to use meta-llama/Meta-Llama-3-70B-Instruct with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use meta-llama/Meta-Llama-3-70B-Instruct with Transformers:

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

pipe = pipeline("text-generation", model="meta-llama/Meta-Llama-3-70B-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-3-70B-Instruct")
model = AutoModelForCausalLM.from_pretrained("meta-llama/Meta-Llama-3-70B-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-3-70B-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-3-70B-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-3-70B-Instruct",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/meta-llama/Meta-Llama-3-70B-Instruct

SGLang

How to use meta-llama/Meta-Llama-3-70B-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-3-70B-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-3-70B-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-3-70B-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-3-70B-Instruct",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use meta-llama/Meta-Llama-3-70B-Instruct with Docker Model Runner:
```
docker model run hf.co/meta-llama/Meta-Llama-3-70B-Instruct
```

Issue with Deprecated Arguments in AutoTrain Advanced for LLaMA 3 (70B) - DPO

#59

by GuusBouwensNL - opened Jul 2, 2024

Discussion

GuusBouwensNL

Jul 2, 2024

Hello Hugging Face Community,

I am currently fine-tuning a LLaMA 3 model using AutoTrain Advanced on the Hugging Face platform. However, I encountered an issue related to deprecated arguments. Here are the details of my setup and the error messages:

Base Model: meta-llama/Meta-Llama-3-70B-Instruct
Task: LLM DPO
Hardware: Local/Space -> 4xL4
Parameter Mode: Full

Error Message:

INFO     | 2024-06-28 12:25:00 | autotrain.trainers.clm.train_clm_dpo:train:57 - Using PEFT, model_ref will be set to None
INFO     | 2024-06-28 12:25:00 | autotrain.trainers.clm.train_clm_dpo:train:78 - model dtype: torch.float16
INFO     | 2024-06-28 12:25:00 | autotrain.trainers.clm.train_clm_dpo:train:95 - creating trainer
/app/env/lib/python3.10/site-packages/huggingface_hub/utils/_deprecation.py:100: FutureWarning: Deprecated argument(s) used in '__init__': max_length, max_prompt_length. Will not be supported from version '1.0.0'.

Deprecated positional argument(s) used in DPOTrainer, please use the DPOConfig to set these arguments instead.
  warnings.warn(message, FutureWarning)
ERROR    | 2024-06-28 12:25:01 | autotrain.trainers.common:wrapper:120 - train has failed due to an exception: Traceback (most recent call last):
  File "/app/env/lib/python3.10/site-packages/autotrain/trainers/common.py", line 117, in wrapper
    return func(*args, **kwargs)
  File "/app/env/lib/python3.10/site-packages/autotrain/trainers/clm/__main__.py", line 38, in train
    train_dpo(config)
  File "/app/env/lib/python3.10/site-packages/autotrain/trainers/clm/train_clm_dpo.py", line 103, in train
    trainer = DPOTrainer(
  File "/app/env/lib/python3.10/site-packages/huggingface_hub/utils/_deprecation.py", line 101, in inner_f
    return f(*args, **kwargs)
  File "/app/env/lib/python3.10/site-packages/trl/trainer/dpo_trainer.py", line 174, in __init__
    if args.model_init_kwargs is None:
AttributeError: 'TrainingArguments' object has no attribute 'model_init_kwargs'

Request for Assistance:

I am seeking guidance on how to properly configure the DPOTrainer using the DPOConfig within the AutoTrain Advanced interface or through a custom configuration file. Specifically, how can I resolve the deprecated argument issue (max_length, max_prompt_length) and the model_init_kwargs attribute error?

Here is a screenshot of my current setup in AutoTrain Advanced:

Thank you in advance for your help!

Best regards,
Guus Bouwens

GuusBouwensNL changed discussion status to closed Jul 5, 2024

GuusBouwensNL changed discussion status to open Jul 5, 2024

charlescumpsty

Aug 13, 2024

@GuusBouwensNL Did you ever find a solution to this problem? I am currently facing the same exact issue with Llama 3.1 8B DPO

yiting

Aug 13, 2024

downgrading trl==0.8.1 works for me

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