uisikdag commited on
Commit
dcc9a7c
1 Parent(s): 7185043

Add yolov5 model card

Browse files
Files changed (1) hide show
  1. README.md +80 -0
README.md ADDED
@@ -0,0 +1,80 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+
2
+ ---
3
+ tags:
4
+ - yolov5
5
+ - yolo
6
+ - vision
7
+ - object-detection
8
+ - pytorch
9
+ library_name: yolov5
10
+ library_version: 7.0.7
11
+ inference: false
12
+
13
+ model-index:
14
+ - name: uisikdag/hardhat
15
+ results:
16
+ - task:
17
+ type: object-detection
18
+
19
+ metrics:
20
+ - type: precision # since mAP@0.5 is not available on hf.co/metrics
21
+ value: 0.9284727818930048 # min: 0.0 - max: 1.0
22
+ name: mAP@0.5
23
+ ---
24
+
25
+ <div align="center">
26
+ <img width="640" alt="uisikdag/hardhat" src="https://huggingface.co/uisikdag/hardhat/resolve/main/sample_visuals.jpg">
27
+ </div>
28
+
29
+ ### How to use
30
+
31
+ - Install [yolov5](https://github.com/fcakyon/yolov5-pip):
32
+
33
+ ```bash
34
+ pip install -U yolov5
35
+ ```
36
+
37
+ - Load model and perform prediction:
38
+
39
+ ```python
40
+ import yolov5
41
+
42
+ # load model
43
+ model = yolov5.load('uisikdag/hardhat')
44
+
45
+ # set model parameters
46
+ model.conf = 0.25 # NMS confidence threshold
47
+ model.iou = 0.45 # NMS IoU threshold
48
+ model.agnostic = False # NMS class-agnostic
49
+ model.multi_label = False # NMS multiple labels per box
50
+ model.max_det = 1000 # maximum number of detections per image
51
+
52
+ # set image
53
+ img = 'https://github.com/ultralytics/yolov5/raw/master/data/images/zidane.jpg'
54
+
55
+ # perform inference
56
+ results = model(img, size=640)
57
+
58
+ # inference with test time augmentation
59
+ results = model(img, augment=True)
60
+
61
+ # parse results
62
+ predictions = results.pred[0]
63
+ boxes = predictions[:, :4] # x1, y1, x2, y2
64
+ scores = predictions[:, 4]
65
+ categories = predictions[:, 5]
66
+
67
+ # show detection bounding boxes on image
68
+ results.show()
69
+
70
+ # save results into "results/" folder
71
+ results.save(save_dir='results/')
72
+ ```
73
+
74
+ - Finetune the model on your custom dataset:
75
+
76
+ ```bash
77
+ yolov5 train --data data.yaml --img 640 --batch 16 --weights uisikdag/hardhat --epochs 10
78
+ ```
79
+
80
+ **More models available at: [awesome-yolov5-models](https://github.com/keremberke/awesome-yolov5-models)**