ONNX
chemistry
AI4Industry commited on
Commit
92c3c2c
·
verified ·
1 Parent(s): da44397

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +9 -6
README.md CHANGED
@@ -15,12 +15,15 @@ tags:
15
  Compared to [MolDet](https://huggingface.co/UniParser/MolDet), **MolDetv2** leverages more manually annotated training data, with further optimizations specifically for reducing molecular false detections and improving bounding box regression, achieving stronger performance with a smaller model.
16
 
17
 
18
- ## [MolDet-General] General molecule structure detection models
19
 
20
  YOLO11-n weights trained on more than 100k human annotated image crops & synthesis molecule images.
21
 
22
  ![image](https://cdn-uploads.huggingface.co/production/uploads/65f7f16fb6941db5c2e7c4bf/iZqZ8rUsD6jacIJr8Hbag.png)
23
 
 
 
 
24
  * 640x640 input resolution
25
  * support handwritten molecules detection
26
  * **multiscale input** (inputs can be single/multiple molecular cutouts, reaction or table cutouts, or single-page PDF images)
@@ -36,12 +39,16 @@ For further usage instructions, please refer to the [official Ultralytics docume
36
 
37
 
38
 
39
- ## [MolDet-Doc] PDF molecule structure detection models
40
 
41
  YOLO11-n weights trained on more than 60k human annotated PDF pages (patents, papers, and books) and 10k synthesis PDF pages with molecule images.
42
 
43
  ![image](https://cdn-uploads.huggingface.co/production/uploads/65f7f16fb6941db5c2e7c4bf/rKZjaZ0EingRtxdIe5Ptz.png)
44
 
 
 
 
 
45
  * 960x960 input resolution
46
  * prefer **single page PDF image** input
47
  * better in small molecule detection
@@ -67,10 +74,6 @@ For further usage instructions, please refer to the [official Ultralytics docume
67
 
68
  Please refer to [MolDet-Bench](https://huggingface.co/datasets/UniParser/MolDet-Bench)
69
 
70
- ![image](https://cdn-uploads.huggingface.co/production/uploads/65f7f16fb6941db5c2e7c4bf/cVEMAL4xpy-y2vgmQ5mR-.png)
71
- ![image](https://cdn-uploads.huggingface.co/production/uploads/65f7f16fb6941db5c2e7c4bf/KHwlqpRZzTOlbOdCG7SyD.png)
72
- ![image](https://cdn-uploads.huggingface.co/production/uploads/65f7f16fb6941db5c2e7c4bf/l-oNYAv3aGbqvGjImYstW.png)
73
-
74
 
75
  ## 📖 Citation
76
 
 
15
  Compared to [MolDet](https://huggingface.co/UniParser/MolDet), **MolDetv2** leverages more manually annotated training data, with further optimizations specifically for reducing molecular false detections and improving bounding box regression, achieving stronger performance with a smaller model.
16
 
17
 
18
+ ## [MolDet-General] universal molecule structure detection
19
 
20
  YOLO11-n weights trained on more than 100k human annotated image crops & synthesis molecule images.
21
 
22
  ![image](https://cdn-uploads.huggingface.co/production/uploads/65f7f16fb6941db5c2e7c4bf/iZqZ8rUsD6jacIJr8Hbag.png)
23
 
24
+ ![image](https://cdn-uploads.huggingface.co/production/uploads/65f7f16fb6941db5c2e7c4bf/cVEMAL4xpy-y2vgmQ5mR-.png)
25
+
26
+ features:
27
  * 640x640 input resolution
28
  * support handwritten molecules detection
29
  * **multiscale input** (inputs can be single/multiple molecular cutouts, reaction or table cutouts, or single-page PDF images)
 
39
 
40
 
41
 
42
+ ## [MolDet-Doc] document molecule structure detection
43
 
44
  YOLO11-n weights trained on more than 60k human annotated PDF pages (patents, papers, and books) and 10k synthesis PDF pages with molecule images.
45
 
46
  ![image](https://cdn-uploads.huggingface.co/production/uploads/65f7f16fb6941db5c2e7c4bf/rKZjaZ0EingRtxdIe5Ptz.png)
47
 
48
+ ![image](https://cdn-uploads.huggingface.co/production/uploads/65f7f16fb6941db5c2e7c4bf/KHwlqpRZzTOlbOdCG7SyD.png)
49
+ ![image](https://cdn-uploads.huggingface.co/production/uploads/65f7f16fb6941db5c2e7c4bf/l-oNYAv3aGbqvGjImYstW.png)
50
+
51
+ features:
52
  * 960x960 input resolution
53
  * prefer **single page PDF image** input
54
  * better in small molecule detection
 
74
 
75
  Please refer to [MolDet-Bench](https://huggingface.co/datasets/UniParser/MolDet-Bench)
76
 
 
 
 
 
77
 
78
  ## 📖 Citation
79