---
license: apache-2.0
datasets:
- HuggingFaceM4/DocumentVQA
language:
- en
library_name: transformers
pipeline_tag: image-text-to-text
---

# Florence-2-finetuned-HuggingFaceM4-DOcumentVQA

This model is a fine-tuned version of [microsoft/Florence-2-base-ft](https://huggingface.co/microsoft/Florence-2-base-ft) on [HuggingFaceM4/DocumentVQA](https://huggingface.co/datasets/HuggingFaceM4/DocumentVQA) dataset.

It is the result of the post [Fine tuning Florence-2](https://maximofn.com/fine-tuning-florence-2/)

It achieves the following results on the evaluation set:
- Loss: 0.7168

## Model description

Florence-2 is an advanced vision foundation model that uses a prompt-based approach to handle a wide range of vision and vision-language tasks. Florence-2 can interpret simple text prompts to perform tasks like captioning, object detection, and segmentation. It leverages our FLD-5B dataset, containing 5.4 billion annotations across 126 million images, to master multi-task learning. The model's sequence-to-sequence architecture enables it to excel in both zero-shot and fine-tuned settings, proving to be a competitive vision foundation model.

He has also been finetuned in the docVQA task.

## Training and evaluation data

This is finetuned on [HuggingFaceM4/DocumentVQA](https://huggingface.co/datasets/HuggingFaceM4/DocumentVQA) dataset.

## Training procedure

### Training hyperparameters

The following hyperparameters were used during training:
- learning_rate: 1e-6
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- optimizer: AdamW with betas=(0.9,0.999) and epsilon=1e-08
- num_epochs: 3

### Training results

| Training Loss | Epoch | Validation Loss |
|:-------------:|:-----:|:---------------:|
| 1.1535        | 1.0   | 0.7698          |
| 0.6530        | 2.0   | 0.7253          |
| 0.5878        | 3.0   | 0.7168          |


### Framework versions

- Transformers 4.43.3
- Pytorch 2.3.1+cu121
- Datasets 2.20.0
- Tokenizers 0.19.1