Instructions to use anon8231489123/vicuna-13b-GPTQ-4bit-128g with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use anon8231489123/vicuna-13b-GPTQ-4bit-128g with Transformers:

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

pipe = pipeline("text-generation", model="anon8231489123/vicuna-13b-GPTQ-4bit-128g")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("anon8231489123/vicuna-13b-GPTQ-4bit-128g")
model = AutoModelForCausalLM.from_pretrained("anon8231489123/vicuna-13b-GPTQ-4bit-128g", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use anon8231489123/vicuna-13b-GPTQ-4bit-128g with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "anon8231489123/vicuna-13b-GPTQ-4bit-128g"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "anon8231489123/vicuna-13b-GPTQ-4bit-128g",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/anon8231489123/vicuna-13b-GPTQ-4bit-128g

SGLang

How to use anon8231489123/vicuna-13b-GPTQ-4bit-128g 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 "anon8231489123/vicuna-13b-GPTQ-4bit-128g" \
    --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": "anon8231489123/vicuna-13b-GPTQ-4bit-128g",
		"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 "anon8231489123/vicuna-13b-GPTQ-4bit-128g" \
        --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": "anon8231489123/vicuna-13b-GPTQ-4bit-128g",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use anon8231489123/vicuna-13b-GPTQ-4bit-128g with Docker Model Runner:
```
docker model run hf.co/anon8231489123/vicuna-13b-GPTQ-4bit-128g
```

Instruction Format

by tarunchand - opened Apr 4, 2023

Discussion

tarunchand

Apr 4, 2023

what is the instruction format prompt used to run this model? Is it the same as the alpaca

anon8231489123

Owner Apr 4, 2023

•

edited Apr 4, 2023

### Human:
### Assistant:

underlines

Apr 5, 2023

i git pulled the latest changes of textgen, added the format to the brand new text-generation-webui\characters\instruction-following\Vicuna.yaml

name: "### Assistant:"
your_name: "### Human:"
context: "Below is an instruction that describes a task. Write a response that appropriately completes the request."

and the model still continues with hallucinated conversations as Human and Assistant. I guess the instructions for fine tuning weren't split and instead trained on multiple questions/answers?

anon8231489123

Owner Apr 7, 2023

Pushed another update to the tokenizer models. May help now. Use latest oobabooga and use the new instruct mode with the vicuna prompt. Should help.

anon8231489123 changed discussion status to closed Apr 7, 2023

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