Instructions to use Jeesup/MUSE-News_Llama-2-7b_graddiff_lr5e-6_lam20_ep5_bestretain with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Jeesup/MUSE-News_Llama-2-7b_graddiff_lr5e-6_lam20_ep5_bestretain with Transformers:

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

pipe = pipeline("text-generation", model="Jeesup/MUSE-News_Llama-2-7b_graddiff_lr5e-6_lam20_ep5_bestretain")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("Jeesup/MUSE-News_Llama-2-7b_graddiff_lr5e-6_lam20_ep5_bestretain")
model = AutoModelForCausalLM.from_pretrained("Jeesup/MUSE-News_Llama-2-7b_graddiff_lr5e-6_lam20_ep5_bestretain", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use Jeesup/MUSE-News_Llama-2-7b_graddiff_lr5e-6_lam20_ep5_bestretain with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Jeesup/MUSE-News_Llama-2-7b_graddiff_lr5e-6_lam20_ep5_bestretain"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Jeesup/MUSE-News_Llama-2-7b_graddiff_lr5e-6_lam20_ep5_bestretain",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/Jeesup/MUSE-News_Llama-2-7b_graddiff_lr5e-6_lam20_ep5_bestretain

SGLang

How to use Jeesup/MUSE-News_Llama-2-7b_graddiff_lr5e-6_lam20_ep5_bestretain 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 "Jeesup/MUSE-News_Llama-2-7b_graddiff_lr5e-6_lam20_ep5_bestretain" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Jeesup/MUSE-News_Llama-2-7b_graddiff_lr5e-6_lam20_ep5_bestretain",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

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 "Jeesup/MUSE-News_Llama-2-7b_graddiff_lr5e-6_lam20_ep5_bestretain" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Jeesup/MUSE-News_Llama-2-7b_graddiff_lr5e-6_lam20_ep5_bestretain",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use Jeesup/MUSE-News_Llama-2-7b_graddiff_lr5e-6_lam20_ep5_bestretain with Docker Model Runner:
```
docker model run hf.co/Jeesup/MUSE-News_Llama-2-7b_graddiff_lr5e-6_lam20_ep5_bestretain
```

Jeesup/MUSE-News_Llama-2-7b_graddiff_lr5e-6_lam20_ep5_bestretain

MUSE unlearning run uploaded by run_muse_settings_to_hf.py.

Corpus: news
Algorithm: graddiff (ga_gdr)
Learning rate: 5e-06
Retain loss coefficient lambda: 20.0
Epochs: 5
NPO beta: not applicable

See muse_run_metadata.json for checkpoint selection details.

Downloads last month: 10

Safetensors

Model size

7B params

Tensor type

BF16

Model tree for Jeesup/MUSE-News_Llama-2-7b_graddiff_lr5e-6_lam20_ep5_bestretain

Base model

muse-bench/MUSE-news_target

Finetuned

(36)

this model