Instructions to use pascal-maker/paligemma_vqav2 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use pascal-maker/paligemma_vqav2 with Transformers:

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

pipe = pipeline("image-text-to-text", model="pascal-maker/paligemma_vqav2")

# Load model directly
from transformers import AutoProcessor, AutoModelForMultimodalLM

processor = AutoProcessor.from_pretrained("pascal-maker/paligemma_vqav2")
model = AutoModelForMultimodalLM.from_pretrained("pascal-maker/paligemma_vqav2", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use pascal-maker/paligemma_vqav2 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "pascal-maker/paligemma_vqav2"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "pascal-maker/paligemma_vqav2",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/pascal-maker/paligemma_vqav2

SGLang

How to use pascal-maker/paligemma_vqav2 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 "pascal-maker/paligemma_vqav2" \
    --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": "pascal-maker/paligemma_vqav2",
		"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 "pascal-maker/paligemma_vqav2" \
        --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": "pascal-maker/paligemma_vqav2",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use pascal-maker/paligemma_vqav2 with Docker Model Runner:
```
docker model run hf.co/pascal-maker/paligemma_vqav2
```

paligemma_vqav2 / config.json

pascal-maker

Training in progress, step 276

985658f verified over 1 year ago

Raw

History Blame Contribute Delete

998 Bytes

	{
	"_name_or_path": "google/paligemma-3b-pt-224",
	"_vocab_size": 257216,
	"architectures": [
	"PaliGemmaForConditionalGeneration"
	],
	"bos_token_id": 2,
	"eos_token_id": 1,
	"hidden_size": 2048,
	"image_token_index": 257152,
	"model_type": "paligemma",
	"pad_token_id": 0,
	"projection_dim": 2048,
	"text_config": {
	"hidden_size": 2048,
	"intermediate_size": 16384,
	"model_type": "gemma",
	"num_attention_heads": 8,
	"num_hidden_layers": 18,
	"num_image_tokens": 256,
	"num_key_value_heads": 1,
	"torch_dtype": "float32",
	"vocab_size": 257216
	},
	"torch_dtype": "bfloat16",
	"transformers_version": "4.47.0.dev0",
	"vision_config": {
	"hidden_size": 1152,
	"intermediate_size": 4304,
	"model_type": "siglip_vision_model",
	"num_attention_heads": 16,
	"num_hidden_layers": 27,
	"num_image_tokens": 256,
	"patch_size": 14,
	"projection_dim": 2048,
	"projector_hidden_act": "gelu_fast",
	"vision_use_head": false
	}
	}