Instructions to use QuixiAI/Llama-3.2-1B-FP8-Dynamic with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use QuixiAI/Llama-3.2-1B-FP8-Dynamic with Transformers:

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

pipe = pipeline("text-generation", model="QuixiAI/Llama-3.2-1B-FP8-Dynamic")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("QuixiAI/Llama-3.2-1B-FP8-Dynamic")
model = AutoModelForCausalLM.from_pretrained("QuixiAI/Llama-3.2-1B-FP8-Dynamic")
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]:]))

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use QuixiAI/Llama-3.2-1B-FP8-Dynamic with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "QuixiAI/Llama-3.2-1B-FP8-Dynamic"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "QuixiAI/Llama-3.2-1B-FP8-Dynamic",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/QuixiAI/Llama-3.2-1B-FP8-Dynamic

SGLang

How to use QuixiAI/Llama-3.2-1B-FP8-Dynamic 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 "QuixiAI/Llama-3.2-1B-FP8-Dynamic" \
    --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": "QuixiAI/Llama-3.2-1B-FP8-Dynamic",
		"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 "QuixiAI/Llama-3.2-1B-FP8-Dynamic" \
        --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": "QuixiAI/Llama-3.2-1B-FP8-Dynamic",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use QuixiAI/Llama-3.2-1B-FP8-Dynamic with Docker Model Runner:
```
docker model run hf.co/QuixiAI/Llama-3.2-1B-FP8-Dynamic
```

ehartford commited on Jan 5

Commit

ddb92be

verified ·

1 Parent(s): 8118f7b

Create README.md

Browse files

Files changed (1) hide show

README.md +29 -0

README.md ADDED Viewed

	@@ -0,0 +1,29 @@

+---
+base_model: meta-llama/Llama-3.2-1B-Instruct
+language:
+- en
+library_name: transformers
+license: llama3.2
+tags:
+- llama-3
+- llama
+- meta
+- facebook
+- transformers
+---
+Quantizing Llama-3.2-1B
+Eric Hartford
+I am creating several quants of Llama-3.1-1B for the purposes of testing vLLM Marlin.
+- https://huggingface.co/QuixiAI/Llama-3.2-1B
+- https://huggingface.co/QuixiAI/Llama-3.2-1B-FP8-Dynamic
+- https://huggingface.co/QuixiAI/Llama-3.2-1B-MXFP4
+- https://huggingface.co/QuixiAI/Llama-3.2-1B-NVFP4A16
+- https://huggingface.co/QuixiAI/Llama-3.2-1B-W4A16-AWQ
+- https://huggingface.co/QuixiAI/Llama-3.2-1B-W4A16-GPTQ
+- https://huggingface.co/QuixiAI/Llama-3.2-1B-W8A16-GPTQ
+The script I used to quant this:
+[quant.py](quant.py)