Instructions to use meta-llama/Llama-3.1-70B-Instruct with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use meta-llama/Llama-3.1-70B-Instruct with Transformers:

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

pipe = pipeline("text-generation", model="meta-llama/Llama-3.1-70B-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("meta-llama/Llama-3.1-70B-Instruct")
model = AutoModelForCausalLM.from_pretrained("meta-llama/Llama-3.1-70B-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=256)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Inference
Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use meta-llama/Llama-3.1-70B-Instruct with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "meta-llama/Llama-3.1-70B-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/Llama-3.1-70B-Instruct",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/meta-llama/Llama-3.1-70B-Instruct

SGLang

How to use meta-llama/Llama-3.1-70B-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/Llama-3.1-70B-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/Llama-3.1-70B-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/Llama-3.1-70B-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/Llama-3.1-70B-Instruct",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use meta-llama/Llama-3.1-70B-Instruct with Docker Model Runner:
```
docker model run hf.co/meta-llama/Llama-3.1-70B-Instruct
```

Issue with Tokenizer when deploying with TGI

#10

by BigBoyAlan - opened Jul 24, 2024

Discussion

BigBoyAlan

Jul 24, 2024

Hey all, I am having an issue with tokenization when trying to deploy Llama 3.1 70B.

I am running this command from CLI. I am running 4x A100 40GB.

sudo docker run --gpus all --shm-size 128g -p 8080:80 -v $volume:/data -v $host_cache_dir:/root/.cache/huggingface -e HUGGINGFACE_TOKEN=$HUGGINGFACE_TOKEN -e RANK=0 -e WORLD_SIZE=1 -e MASTER_ADDR=localhost -e MASTER_PORT=29500 -e CUDA_VISIBLE_DEVICES=0,1,2,3 ghcr.io/huggingface/text-generation-inference:2.2.0 --model-id $model --sharded true --quantize eetq --cuda-memory-fraction 0.6

Deployment Arguments:
model_id: "meta-llama/Meta-Llama-3.1-70B-Instruct", revision: None, validation_workers: 2, sharded: Some( true, ), num_shard: None, quantize: Some( Eetq, ), speculate: None, dtype: None, trust_remote_code: false, max_concurrent_requests: 128, max_best_of: 2, max_stop_sequences: 4, max_top_n_tokens: 5, max_input_tokens: None, max_input_length: None, max_total_tokens: None, waiting_served_ratio: 0.3, max_batch_prefill_tokens: None, max_batch_total_tokens: None, max_waiting_tokens: 20, max_batch_size: None, cuda_graphs: None, hostname: "0541246362b8", port: 80, shard_uds_path: "/tmp/text-generation-server", master_addr: "localhost", master_port: 29500, huggingface_hub_cache: Some( "/data", ), weights_cache_override: None, disable_custom_kernels: false, cuda_memory_fraction: 0.6, rope_scaling: None, rope_factor: None, json_output: false, otlp_endpoint: None, otlp_service_name: "text-generation-inference.router", cors_allow_origin: [], watermark_gamma: None, watermark_delta: None, ngrok: false, ngrok_authtoken: None, ngrok_edge: None, tokenizer_config_path: None, disable_grammar_support: false, env: false, max_client_batch_size: 4, lora_adapters: None, disable_usage_stats: false, disable_crash_reports: false, }

This runs properly, but whenever I try to generate a response from /generate, I am getting this response:

data: {"index":1,"token":{"id":52107,"text":"\\\\","logprob":0.0,"special":false},"top_tokens":[{"id":52107,"text":"\\\\","logprob":0.0,"special":false}],"generated_text":null,"details":null}

data: {"index":2,"token":{"id":52107,"text":"\\\\","logprob":0.0,"special":false},"top_tokens":[{"id":52107,"text":"\\\\","logprob":0.0,"special":false}],"generated_text":null,"details":null}

...

my initial thoughts were issues with the tokenizer, so I tested the /tokenize endpoint:

[
{
"id": 5159,
"text": "",
"start": 0,
"stop": 0
},
{
"id": 836,
"text": "",
"start": 2,
"stop": 2
},
{
"id": 374,
"text": "",
"start": 7,
"stop": 7
},
{
"id": 78118,
"text": "",
"start": 10,
"stop": 10
},
{
"id": 323,
"text": "",
"start": 18,
"stop": 18
},

okay, so the fact that the text was not popping up was that there must have been an issue with the tokenizer. I then downloaded the tokenizer.json from the model files and manually set it when running the model using the --tokenizer-config-path, but it still did not fix the issue. Has anyone ran into a similar issue?

any help would be greatly appreciated.

gauravgary

Aug 8, 2024

Yes I can confirm this is an issue. Though the file you probably need to specify is tokenizer_config.json, however TGI is not using the file for some reason.

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