|
--- |
|
datasets: |
|
- agentsea/wave-ui |
|
language: |
|
- en |
|
library_name: transformers |
|
--- |
|
|
|
# Paligemma WaveUI |
|
|
|
|
|
Transformers [PaliGemma 3B 448-res weights](https://huggingface.co/google/paligemma-3b-pt-448), fine-tuned on the [WaveUI](https://huggingface.co/datasets/agentsea/wave-ui) dataset for object-detection. |
|
|
|
## Model Details |
|
|
|
### Model Description |
|
|
|
This fine-tune was done atop of the [Paligemma 448 Widgetcap](https://huggingface.co/google/paligemma-3b-ft-widgetcap-448) model, using the [WaveUI](https://huggingface.co/datasets/agentsea/wave-ui) dataset, which contains ~80k examples of labeled UI elements. |
|
|
|
The fine-tune was done for the object detection task. Specifically, this model aims to perform well at UI element detection, as part of a wider effort to enable our open-source toolkit for building agents at [AgentSea](https://www.agentsea.ai/). |
|
|
|
- **Developed by:** https://agentsea.ai/ |
|
- **Language(s) (NLP):** en |
|
- **Finetuned from model:** https://huggingface.co/google/paligemma-3b-ft-widgetcap-448 |
|
|
|
### Demo |
|
|
|
You can find a **demo** for this model [here](https://huggingface.co/spaces/agentsea/paligemma-waveui). |
|
|
|
## Notes |
|
|
|
- The only task used in the fine-tune was the object detection task, so it might not perform well in other types of tasks. |
|
|
|
## Usage |
|
|
|
To start using this model, run the following: |
|
|
|
```python |
|
from transformers import AutoProcessor, PaliGemmaForConditionalGeneration |
|
|
|
model = PaliGemmaForConditionalGeneration.from_pretrained("agentsea/paligemma-3b-ft-widgetcap-waveui-448").eval() |
|
processor = AutoProcessor.from_pretrained("google/paligemma-3b-pt-448") |
|
``` |
|
|
|
## Data |
|
|
|
We used the [WaveUI](https://huggingface.co/datasets/agentsea/wave-ui) dataset for this fine-tune. Before using it, we preprocessed the data to use the Paligemma bounding-box format. |
|
|
|
|
|
## Evaluation |
|
|
|
We calculated the mean IoU over 1024 examples of the test set using 3 different closed-source models: Gemini 1.5 Pro, Claude 3.5 Sonnet and GPT 4o. We also ran this same calculation using the PaliGemma WaveUI fine-tunes. We obtained the following values: |
|
|
|
- Gemini 1.5 Pro: 0.12 |
|
- Claude 3.5 Sonnet: 0.05 |
|
- GPT 4o: 0.05 |
|
- **PaliGemma Widgetcap+WaveUI 448: 0.40** |
|
- PaliGemma WaveUI 896: 0.49 |
