Instructions to use google/paligemma-3b-mix-224 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use google/paligemma-3b-mix-224 with Transformers:

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

pipe = pipeline("image-text-to-text", model="google/paligemma-3b-mix-224")

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoProcessor, AutoModelForMultimodalLM

processor = AutoProcessor.from_pretrained("google/paligemma-3b-mix-224")
model = AutoModelForMultimodalLM.from_pretrained("google/paligemma-3b-mix-224", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use google/paligemma-3b-mix-224 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "google/paligemma-3b-mix-224"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "google/paligemma-3b-mix-224",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/google/paligemma-3b-mix-224

SGLang

How to use google/paligemma-3b-mix-224 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/paligemma-3b-mix-224" \
    --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/paligemma-3b-mix-224",
		"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/paligemma-3b-mix-224" \
        --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/paligemma-3b-mix-224",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use google/paligemma-3b-mix-224 with Docker Model Runner:
```
docker model run hf.co/google/paligemma-3b-mix-224
```

Inference via sagemaker endpoint

#12

by shum123 - opened Jul 31, 2024

Discussion

shum123

Jul 31, 2024

Hi! How do i call this for inference with an image via a sagemaker endpoint?

send request

predictor.predict({
"inputs": "My name is Clara and I am", "parameters":parameters
})
This works, but how do i include an image in the input?

I tried this:
image_path ='./images/powercord.jpg'
with open(image_path, "rb") as f:
image = base64.b64encode(f.read()).decode("utf-8")

payload = { "inputs" :{
"prompt": "detect powercord",
"image": image
}}

query_response = predictor.predict(data=payload)

But im getting this error - An error occurred (ModelError) when calling the InvokeEndpoint operation: Received client error (422) from primary with message "Failed to deserialize the JSON body into the target type: inputs: invalid type: map, expected a string at line 1 column 11".

lkv

Jan 30, 2025

Hi @shum123 ,

The model is expecting a JSON format where the inputs are passed in a way that it can deserialize into the expected structure. You're including both the prompt and the image in the inputs dictionary. You should pass the entire payload as a valid JSON string (json.dumps(payload)) in your request. Kindly use the below code.

payload = { "inputs": json.dumps({ "prompt": "detect powercord", "image": image }) }

Thank you.

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