Implementation of vision-language models including PaliGemma.
PaliGemma is a vision-language model for image understanding and text generation.
-
Download the model:
python PaliGemma/download_model.py -
Install dependencies (PyTorch, transformers, PIL, etc.)
Run inference:
python PaliGemma/inference.py --image <path> --prompt <text>
Or use the shell script:
bash PaliGemma/inference.sh
Fine-tune the model:
python PaliGemma/finetune.py --train_data <json_file>
Or use the shell script:
bash PaliGemma/finetune.sh
inference.py- Model inference scriptfinetune.py- Fine-tuning scriptdownload_model.py- Model download utilitypaligemma.py- PaliGemma model implementationgemma.py- Gemma language model componentsprocessing.py- Image and text processing utilities