Instructions to use NousResearch/GPT4-x-Vicuna-13b-fp16 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use NousResearch/GPT4-x-Vicuna-13b-fp16 with Transformers:

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

pipe = pipeline("text-generation", model="NousResearch/GPT4-x-Vicuna-13b-fp16")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("NousResearch/GPT4-x-Vicuna-13b-fp16")
model = AutoModelForCausalLM.from_pretrained("NousResearch/GPT4-x-Vicuna-13b-fp16", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use NousResearch/GPT4-x-Vicuna-13b-fp16 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "NousResearch/GPT4-x-Vicuna-13b-fp16"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "NousResearch/GPT4-x-Vicuna-13b-fp16",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/NousResearch/GPT4-x-Vicuna-13b-fp16

SGLang

How to use NousResearch/GPT4-x-Vicuna-13b-fp16 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 "NousResearch/GPT4-x-Vicuna-13b-fp16" \
    --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": "NousResearch/GPT4-x-Vicuna-13b-fp16",
		"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 "NousResearch/GPT4-x-Vicuna-13b-fp16" \
        --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": "NousResearch/GPT4-x-Vicuna-13b-fp16",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use NousResearch/GPT4-x-Vicuna-13b-fp16 with Docker Model Runner:
```
docker model run hf.co/NousResearch/GPT4-x-Vicuna-13b-fp16
```

As a base model used https://huggingface.co/eachadea/vicuna-13b-1.1

Finetuned on Teknium's GPTeacher dataset, unreleased Roleplay v2 dataset, GPT-4-LLM dataset Uncensored, WizardLM Uncensored and Nous Research Instruct Dataset

Approx 180k instructions, all from GPT-4, all cleaned of any OpenAI censorship/"As an AI Language Model" etc.

Base model still has OpenAI censorship. Soon, a new version will be released with cleaned vicuna from https://huggingface.co/datasets/anon8231489123/ShareGPT_Vicuna_unfiltere

Trained on 8 A100-80GB GPUs for 5 epochs following Alpaca deepspeed training code.

Nous Research Instruct Dataset will be released soon.

Prompt format is Alpaca:

### Instruction:

### Response:

### Instruction:

### Input:

### Response:

GPTeacher, Roleplay v2 by https://huggingface.co/teknium

Wizard LM by https://github.com/nlpxucan

Nous Research Instruct Dataset by https://huggingface.co/karan4d and https://huggingface.co/huemin

Compute provided by our project sponsor https://redmond.ai/

Downloads last month: 136

Model tree for NousResearch/GPT4-x-Vicuna-13b-fp16

Adapters

127 models

NousResearch
/

GPT4-x-Vicuna-13b-fp16

Model tree for NousResearch/GPT4-x-Vicuna-13b-fp16

Space using NousResearch/GPT4-x-Vicuna-13b-fp16 1