Instructions to use ebrukilic/paligemma2_vizwiz_ft2 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use ebrukilic/paligemma2_vizwiz_ft2 with Transformers:

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

pipe = pipeline("image-text-to-text", model="ebrukilic/paligemma2_vizwiz_ft2")

# Load model directly
from transformers import AutoProcessor, AutoModelForMultimodalLM

processor = AutoProcessor.from_pretrained("ebrukilic/paligemma2_vizwiz_ft2")
model = AutoModelForMultimodalLM.from_pretrained("ebrukilic/paligemma2_vizwiz_ft2", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use ebrukilic/paligemma2_vizwiz_ft2 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "ebrukilic/paligemma2_vizwiz_ft2"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "ebrukilic/paligemma2_vizwiz_ft2",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/ebrukilic/paligemma2_vizwiz_ft2

SGLang

How to use ebrukilic/paligemma2_vizwiz_ft2 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 "ebrukilic/paligemma2_vizwiz_ft2" \
    --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": "ebrukilic/paligemma2_vizwiz_ft2",
		"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 "ebrukilic/paligemma2_vizwiz_ft2" \
        --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": "ebrukilic/paligemma2_vizwiz_ft2",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use ebrukilic/paligemma2_vizwiz_ft2 with Docker Model Runner:
```
docker model run hf.co/ebrukilic/paligemma2_vizwiz_ft2
```

paligemma2_vizwiz_ft2

This model is a fine-tuned version of google/paligemma2-3b-pt-448 on the None dataset. It achieves the following results on the evaluation set:

Loss: 0.6072

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

learning_rate: 1e-05
train_batch_size: 1
eval_batch_size: 2
seed: 42
gradient_accumulation_steps: 4
total_train_batch_size: 4
optimizer: Use OptimizerNames.PAGED_ADAMW_8BIT with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
lr_scheduler_type: linear
training_steps: 4000

Training results

Training Loss	Epoch	Step	Validation Loss
0.7985	0.1308	500	0.7325
0.7331	0.2616	1000	0.7017
0.6924	0.3924	1500	0.6753
0.6626	0.5232	2000	0.6491
0.7161	0.6540	2500	0.6260
0.6387	0.7848	3000	0.6122
0.4991	0.9156	3500	0.6080
0.3359	1.0463	4000	0.6072