Instructions to use compressed-llm/vicuna-13b-v1.3-wanda with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use compressed-llm/vicuna-13b-v1.3-wanda with Transformers:

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

pipe = pipeline("text-generation", model="compressed-llm/vicuna-13b-v1.3-wanda")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("compressed-llm/vicuna-13b-v1.3-wanda")
model = AutoModelForCausalLM.from_pretrained("compressed-llm/vicuna-13b-v1.3-wanda")

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use compressed-llm/vicuna-13b-v1.3-wanda with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "compressed-llm/vicuna-13b-v1.3-wanda"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "compressed-llm/vicuna-13b-v1.3-wanda",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/compressed-llm/vicuna-13b-v1.3-wanda

SGLang

How to use compressed-llm/vicuna-13b-v1.3-wanda 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 "compressed-llm/vicuna-13b-v1.3-wanda" \
    --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": "compressed-llm/vicuna-13b-v1.3-wanda",
		"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 "compressed-llm/vicuna-13b-v1.3-wanda" \
        --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": "compressed-llm/vicuna-13b-v1.3-wanda",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use compressed-llm/vicuna-13b-v1.3-wanda with Docker Model Runner:
```
docker model run hf.co/compressed-llm/vicuna-13b-v1.3-wanda
```

YAML Metadata Warning:empty or missing yaml metadata in repo card

Check out the documentation for more information.

from transformers import AutoTokenizer, AutoModelForCausalLM, AutoConfig
import torch

model_path = 'efficient-llm/vicuna-13b-v1.3-wanda'
config = AutoConfig.from_pretrained(model_path, revision='0.5_2to4', trust_remote_code=True)
enc = AutoTokenizer.from_pretrained('lmsys/vicuna-13b-v1.3', trust_remote_code=True)
kwargs = {"torch_dtype": torch.float16, "low_cpu_mem_usage": True}
model = AutoModelForCausalLM.from_pretrained(
    model_path, config=config, trust_remote_code=True, device_map='auto', revision='0.5_2to4', **kwargs)

model.eval()
input_ids = enc('How are you today?', return_tensors='pt').input_ids.to('cuda')
outputs = model.generate(input_ids=input_ids, max_length=128)
print(enc.decode(outputs[0]))

Downloads last month: 3