Which base paligemma was used?
#1
by
theform1234
- opened
All of the base paligemma models I found are all specific to the size of the images used to train them. I'm wondering if there are any specific size considerations that should be taken into account for inference using these fine tuned models linked to the base paligemma used here?? The base paligemmas were trained on one of the following image sizes: 224x224, 448x448, or 896x896
I used 224 models for all of my fine tunes which are default one in fine tuning google colab notebook.
https://www.kaggle.com/code/nilaychauhan/fine-tune-paligemma-with-jax