README.md
1.7 KB · 76 lines · markdown Raw
1 ---
2 tags:
3 - ultralyticsplus
4 - yolov8
5 - ultralytics
6 - yolo
7 - vision
8 - object-detection
9 - pytorch
10 - awesome-yolov8-models
11 library_name: ultralytics
12 library_version: 8.0.21
13 inference: false
14 datasets:
15 - keremberke/table-extraction
16 model-index:
17 - name: keremberke/yolov8m-table-extraction
18 results:
19 - task:
20 type: object-detection
21 dataset:
22 type: keremberke/table-extraction
23 name: table-extraction
24 split: validation
25 metrics:
26 - type: precision
27 value: 0.95194
28 name: mAP@0.5(box)
29 license: agpl-3.0
30 ---
31
32 <div align="center">
33 <img width="640" alt="keremberke/yolov8m-table-extraction" src="https://huggingface.co/keremberke/yolov8m-table-extraction/resolve/main/thumbnail.jpg">
34 </div>
35
36 ### Supported Labels
37
38 ```
39 ['bordered', 'borderless']
40 ```
41
42 ### How to use
43
44 - Install [ultralyticsplus](https://github.com/fcakyon/ultralyticsplus):
45
46 ```bash
47 pip install ultralyticsplus==0.0.23 ultralytics==8.0.21
48 ```
49
50 - Load model and perform prediction:
51
52 ```python
53 from ultralyticsplus import YOLO, render_result
54
55 # load model
56 model = YOLO('keremberke/yolov8m-table-extraction')
57
58 # set model parameters
59 model.overrides['conf'] = 0.25 # NMS confidence threshold
60 model.overrides['iou'] = 0.45 # NMS IoU threshold
61 model.overrides['agnostic_nms'] = False # NMS class-agnostic
62 model.overrides['max_det'] = 1000 # maximum number of detections per image
63
64 # set image
65 image = 'https://github.com/ultralytics/yolov5/raw/master/data/images/zidane.jpg'
66
67 # perform inference
68 results = model.predict(image)
69
70 # observe results
71 print(results[0].boxes)
72 render = render_result(model=model, image=image, result=results[0])
73 render.show()
74 ```
75
76 **More models available at: [awesome-yolov8-models](https://yolov8.xyz)**