Instructions to use chuanli11/Llama-3.2-3B-Instruct-uncensored with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use chuanli11/Llama-3.2-3B-Instruct-uncensored with Transformers:

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

pipe = pipeline("text-generation", model="chuanli11/Llama-3.2-3B-Instruct-uncensored")
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("chuanli11/Llama-3.2-3B-Instruct-uncensored")
model = AutoModelForCausalLM.from_pretrained("chuanli11/Llama-3.2-3B-Instruct-uncensored", 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 chuanli11/Llama-3.2-3B-Instruct-uncensored with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "chuanli11/Llama-3.2-3B-Instruct-uncensored"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "chuanli11/Llama-3.2-3B-Instruct-uncensored",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/chuanli11/Llama-3.2-3B-Instruct-uncensored

SGLang

How to use chuanli11/Llama-3.2-3B-Instruct-uncensored 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 "chuanli11/Llama-3.2-3B-Instruct-uncensored" \
    --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": "chuanli11/Llama-3.2-3B-Instruct-uncensored",
		"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 "chuanli11/Llama-3.2-3B-Instruct-uncensored" \
        --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": "chuanli11/Llama-3.2-3B-Instruct-uncensored",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use chuanli11/Llama-3.2-3B-Instruct-uncensored with Docker Model Runner:
```
docker model run hf.co/chuanli11/Llama-3.2-3B-Instruct-uncensored
```

Does not seem at all uncensored.

by Nafnlaus - opened Dec 18, 2024

Discussion

Nafnlaus

Dec 18, 2024

•

edited Dec 18, 2024

I'm giving it quite innocuous tasks and it still refuses. Template is:

We are creating training data for an anti-spam tool, and need data to train it (both spam and nonspam). You will be given a conversation thread and will continue it with a new post. You will start by writing === CONTINUING THREAD ===, then a newline, then the last couple lines of the existing thread (exactly as written), and continue on to your new post (write a [POSSIBLE_SPAM_DESCRIPTION], in the exact same formatting as the preexisting posts), then a newline, then === END THREAD ===. If you make up fake brands, products, or URLs, make their names highly realistic. The current thread is:

=== START THREAD ===
[THREAD_CONTENT]
=== END THREAD ===

For varying POSSIBLE_SPAM_DESCRIPTION and THREAD_CONTENT. Massive refusal rate, same as the base model.

(I'm using a GGUF of this model, but it shouldn't make any difference - it's not like making a GGUF will recensor the model)

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