Created README
Browse files
README.md
ADDED
|
@@ -0,0 +1,23 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
thumbnail: https://huggingface.co/front/thumbnails/lladoc.png
|
| 3 |
+
language:
|
| 4 |
+
- en
|
| 5 |
+
license: cc-by-4.0
|
| 6 |
+
tags:
|
| 7 |
+
- transformers
|
| 8 |
+
datasets:
|
| 9 |
+
- idocvqa
|
| 10 |
+
metrics:
|
| 11 |
+
- accuracy
|
| 12 |
+
---
|
| 13 |
+
|
| 14 |
+
## LLaDoc (Large Language and Document) model
|
| 15 |
+
|
| 16 |
+
This is a fine-tuned model of LLaVA1.5 (7B) on the iDocVQA dataset. It is intended to be used as a multimodal system.
|
| 17 |
+
The dataset it's trained on is limited in scope, as it covers only certain domains.
|
| 18 |
+
|
| 19 |
+
The accuracy achieved on the validation set is 29.58%.
|
| 20 |
+
|
| 21 |
+
Please find the information about preprocessing, training and full details of the LLaVA model in the [original link](https://llava-vl.github.io/)
|
| 22 |
+
|
| 23 |
+
The paper for this work is available on arXiv: [https://arxiv.org/abs/2402.00453](https://arxiv.org/abs/2402.00453)
|