Instructions to use espressor/meta-llama.Llama-3.2-1B-Instruct_W4A16 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use espressor/meta-llama.Llama-3.2-1B-Instruct_W4A16 with Transformers:

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

pipe = pipeline("text-generation", model="espressor/meta-llama.Llama-3.2-1B-Instruct_W4A16")
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("espressor/meta-llama.Llama-3.2-1B-Instruct_W4A16")
model = AutoModelForCausalLM.from_pretrained("espressor/meta-llama.Llama-3.2-1B-Instruct_W4A16", 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]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use espressor/meta-llama.Llama-3.2-1B-Instruct_W4A16 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "espressor/meta-llama.Llama-3.2-1B-Instruct_W4A16"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "espressor/meta-llama.Llama-3.2-1B-Instruct_W4A16",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/espressor/meta-llama.Llama-3.2-1B-Instruct_W4A16

SGLang

How to use espressor/meta-llama.Llama-3.2-1B-Instruct_W4A16 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 "espressor/meta-llama.Llama-3.2-1B-Instruct_W4A16" \
    --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": "espressor/meta-llama.Llama-3.2-1B-Instruct_W4A16",
		"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 "espressor/meta-llama.Llama-3.2-1B-Instruct_W4A16" \
        --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": "espressor/meta-llama.Llama-3.2-1B-Instruct_W4A16",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use espressor/meta-llama.Llama-3.2-1B-Instruct_W4A16 with Docker Model Runner:
```
docker model run hf.co/espressor/meta-llama.Llama-3.2-1B-Instruct_W4A16
```

meta-llama/Llama-3.2-1B-Instruct - W4A16 Compression

This is a compressed model using llmcompressor.

Compression Configuration

Base Model: meta-llama/Llama-3.2-1B-Instruct
Compression Scheme: W4A16
Dataset: HuggingFaceH4/ultrachat_200k
Dataset Split: train_sft
Number of Samples: 512
Preprocessor: chat
Maximum Sequence Length: 131072

Sample Output

Prompt:

<|begin_of_text|><|start_header_id|>system<|end_header_id|>

Cutting Knowledge Date: December 2023
Today Date: 28 Nov 2024

<|eot_id|><|start_header_id|>user<|end_header_id|>

Who is Alan Turing?<|eot_id|>

Output:

<|begin_of_text|><|begin_of_text|><|start_header_id|>system<|end_header_id|>

Cutting Knowledge Date: December 2023
Today Date: 28 Nov 2024

<|eot_id|><|start_header_id|>user<|end_header_id|>

Who is Alan Turing?<|eot_id|><|start_header_id|>assistant<|end_header_id|>

Alan Turing (1912-1954) was a British mathematician, computer scientist, logician, and philosopher. He is widely recognized for his groundbreaking work in computer science, artificial intelligence, and cryptography. He is often referred to as the "father of computer science" and is considered one of the most influential figures in the development of modern computing.

Turing's most notable contributions include:

1. **The Turing Machine**: In 1936, Turing proposed a theoretical model of a machine that could simulate the behavior of a human thinker, which laid the foundation for modern computer architecture.
2. **Codebreaking during

Evaluation

Downloads last month: 71

Safetensors

Model size

2B params

Tensor type

I64

I32

BF16

Model tree for espressor/meta-llama.Llama-3.2-1B-Instruct_W4A16

Base model

meta-llama/Llama-3.2-1B-Instruct

Quantized

(435)

this model

espressor
/

meta-llama.Llama-3.2-1B-Instruct_W4A16