Update README.md
Browse files
README.md
CHANGED
|
@@ -1,3 +1,9 @@
|
|
| 1 |
-
---
|
| 2 |
-
license: apache-2.0
|
| 3 |
-
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
---
|
| 4 |
+
|
| 5 |
+
Our dataset is built upon the Uground, Jedi, and additional public paper-style academic document screenshots.
|
| 6 |
+
|
| 7 |
+
**NOTE**: Before you use this dataset, make sure you understand the logic of absolute coordinates and [image processor](https://github.com/QwenLM/Qwen2.5-VL/blob/d2240f11656bfe404b9ba56db4e51cd09f522ff1/qwen-vl-utils/src/qwen_vl_utils/vision_process.py#L60) for [Qwen2.5-VL](https://arxiv.org/abs/2502.13923).
|
| 8 |
+
This dataset is set with the image processor max tokens to be 2700, a.k.a max_pixels=2700x14x14x2x2 , the coordinates were resized to be smaller and you have to resize the image as well within max_pixels=2700x14x14x2x2 via image processor to make them align.
|
| 9 |
+
Make sure you also follow it in your training procedure, otherwise the performance will not be as expected.
|