Instructions to use wojtab/llava-13b-v0-4bit-128g with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use wojtab/llava-13b-v0-4bit-128g with Transformers:

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

pipe = pipeline("text-generation", model="wojtab/llava-13b-v0-4bit-128g")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("wojtab/llava-13b-v0-4bit-128g")
model = AutoModelForCausalLM.from_pretrained("wojtab/llava-13b-v0-4bit-128g")

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use wojtab/llava-13b-v0-4bit-128g with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "wojtab/llava-13b-v0-4bit-128g"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "wojtab/llava-13b-v0-4bit-128g",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/wojtab/llava-13b-v0-4bit-128g

SGLang

How to use wojtab/llava-13b-v0-4bit-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 "wojtab/llava-13b-v0-4bit-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": "wojtab/llava-13b-v0-4bit-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 "wojtab/llava-13b-v0-4bit-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": "wojtab/llava-13b-v0-4bit-128g",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use wojtab/llava-13b-v0-4bit-128g with Docker Model Runner:
```
docker model run hf.co/wojtab/llava-13b-v0-4bit-128g
```

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

Check out the documentation for more information.

4-bit quant of llama part of llava https://github.com/haotian-liu/LLaVA https://huggingface.co/liuhaotian/LLaVA-13b-delta-v0

quantized by:

CUDA_VISIBLE_DEVICES=0 python llama.py /workspace/LLaVA-13B-v0/ c4 --wbits 4 --true-sequential --groupsize 128 --save_safetensors llava-13b-v0-4bit-128g.safetensors

on https://github.com/oobabooga/GPTQ-for-LLaMa CUDA branch of GPTQ (commit 57a2629)

YOU CAN NOW RUN IT IN TEXT-GENERATION-WEBUI with llava extension (see: https://github.com/oobabooga/text-generation-webui/tree/main/extensions/llava)

license: other

Downloads last month: 215

wojtab
/

llava-13b-v0-4bit-128g

license: other

Spaces using wojtab/llava-13b-v0-4bit-128g 3