Instructions to use google/paligemma2-28b-pt-896 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use google/paligemma2-28b-pt-896 with Transformers:

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

pipe = pipeline("image-text-to-text", model="google/paligemma2-28b-pt-896")

# Load model directly
from transformers import AutoProcessor, AutoModelForMultimodalLM

processor = AutoProcessor.from_pretrained("google/paligemma2-28b-pt-896")
model = AutoModelForMultimodalLM.from_pretrained("google/paligemma2-28b-pt-896", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use google/paligemma2-28b-pt-896 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "google/paligemma2-28b-pt-896"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "google/paligemma2-28b-pt-896",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/google/paligemma2-28b-pt-896

SGLang

How to use google/paligemma2-28b-pt-896 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 "google/paligemma2-28b-pt-896" \
    --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": "google/paligemma2-28b-pt-896",
		"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 "google/paligemma2-28b-pt-896" \
        --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": "google/paligemma2-28b-pt-896",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use google/paligemma2-28b-pt-896 with Docker Model Runner:
```
docker model run hf.co/google/paligemma2-28b-pt-896
```

Gated Repo Access

by venkai - opened Dec 23, 2024

Discussion

venkai

Dec 23, 2024

Hi,
I requested access to the Gemma2 family of models on December 18, but I haven't received access yet. I am mainly intersted in trying out the paligemma2 family of models trained on high-res images as well as the one finetuned on Docci. Can someone from the Gemma team help out with the access request please.

GopiUppari

Google org Dec 24, 2024

Hi @venkai ,

I reproduced the issue, for now to get instant approval could you please visit the Kaggle page for Gemma and request access to the model weights through Hugging Face. For more reference, please refer to this blog

Thank you.

venkai

Dec 25, 2024

Hi @GopiUppari ,
I just requested access on kaggle:

On huggingface, it still shows that the access request is pending approval:

venkai

Dec 26, 2024

@GopiUppari it works now :-) ! It seems there was some trouble linking my HF to kaggle account. Have access now :thumbsup

venkai changed discussion status to closed Dec 26, 2024

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