Introduction
NeuralMarker offers a variety of Pretrained Deep Learning Models for Rectangle Category shape from different available frameworks like TensorFlow, Yolo, and PyTorch.
These available Pretrained Deep Learning Models profoundly use for Object Detection tasks to detect objects such as Person, Bicycle, Face,s and many more.
This list of models at hand is not static, as more State-Of-The-Art models in the Neural Network Library of NeuralMarker will be added, then accordingly the below list will be updated too:
Pretrained Models
1. Centernet_COCO
Model Name: Centernet
Network Backbone: Hourglass-104
DLA-34
ResNet-101
ResNet-18
No of Categories: 80
Dataset Used: MS COCO
2. YOLOv3_COCO
Model Name: YOLOv3
Network Backbone: Darknet-53
No of Categories: 80
Dataset Used: MS COCO
3. Efficientdet-d0_COCO
Model Name: Efficientdet-d0
Network Backbone: ImageNet-pretrained EfficientNets
No of Categories: 80
Dataset Used: MS COCO
4. Efficientdet-d3_COCO
Model Name: Efficientdet-d3
Network Backbone: ImageNet-pretrained EfficientNets
No of Categories: 80
Dataset Used: MS COCO
5. MTCNN_widerface
Model Name: Multi-task CNN
No of Categories: 1 (Face)
Dataset Used: WIDER FACE dataset
6. YOLOv5s_COCO
Model Name: YOLOv5s
No of Categories: 80
Dataset Used: MS COCO
7. YOLOv5x_COCO
Model Name: YOLOv5x
No of Categories: 80
Dataset Used: MS COCO
8. Detr-Detection
Model Name: DEtection TRansformer
Network Backbone: R50
No of Categories: 80
Dataset Used: MS COCO
9. Faster R-CNN-mmdet-R50-FPN-1x
Model Library: mmdet
Model Architecture: Faster R-CNN
Network Backbone: R-50-FPN
lr schd: 1x
No of Categories: 80
Dataset Used: MS COCO
10. Faster R-CNN-mmdet-R50-FPN-2x
Model Library: mmdet
Model Architecture: Faster R-CNN
Network Backbone: R-50-FPN
lr schd: 2x
No of Categories: 80
Dataset Used: MS COCO
11. mmdet-RetinaNet-R50-FPN-1x
Model Library : mmdet
Model Architecture : RetinaNet
Network Backbone: R-50-FPN
lr schd: 1x
No of Categories: 80
Dataset Used: MS COCO
12. mmdet-RetinaNet-R50-FPN-2x
Model Library : mmdet
Model Architecture : RetinaNet
Network Backbone: R-50-FPN
lr schd: 2x
No of Categories: 80
Dataset Used: MS COCO
13. Faster R-CNN-Detectron2-R50-FPN
Model Library: Detectron2
Model Architecture: Faster R-CNN
Network Backbone: R-50-FPN
No of Categories: 80
Dataset Used: MS COCO
14. Faster R-CNN-Detectron2-R101-FPN
Model Library: Detectron2
Model Architecture: Faster R-CNN
Network Backbone: R-101-FPN
No of Categories: 80
Dataset Used: MS COCO
15. Faster R-CNN-Detectron2-X101-FPN
Model Library: Detectron2
Model Architecture: Faster R-CNN
Network Backbone: X-101-FPN
No of Categories: 80
Dataset Used: MS COCO
16. RetinaNet-R50
Model Library : Detectron2
Model Architecture : RetinaNet
Network Backbone: R-50
lr schd: 1x
No of Categories: 80
Dataset Used: MS COCO
17. RetinaNet-R50_3x
Model Library : Detectron2
Model Architecture : RetinaNet
Network Backbone: R-50
lr schd: 3x
No of Categories: 80
Dataset Used: MS COCO
18. RetinaNet-R101
Model Library : Detectron2
Model Architecture : RetinaNet
Network Backbone: R-101
lr schd: 3x
No of Categories: 80
Dataset Used: MS COCO
19. TFlite-object-detection-v2-ssd_resnet101_v1_fpn_640x640_coco
Model Architecture : SSD FPN object detection
Network Backbone: R-101-FPN
No of Categories: 80
Dataset Used: MS COCO
20. TFlite-object-detection-v2-ssd_resnet101_v1_fpn_1024x1024_coco
Model Architecture: SSD FPN object detection
Network Backbone: R-101-FPN
No of Categories: 80
Dataset Used: MS COCO
21. TFlite-object-detection-v2-ssd_resnet152_v1_fpn_640x640_coco
Model Architecture: SSD FPN object detection
Network Backbone: R-152-FPN
No of Categories: 80
Dataset Used: MS COCO
22. TFlite-object-detection-v2-ssd_resnet152_v1_fpn_1024x1024_coco
Model Architecture: SSD FPN object detection
Network Backbone: R-152-FPN
No of Categories: 80
Dataset Used: MS COCO
23. TFlite-object-detection-v2-ssd_resnet50_v1_fpn_640x640_coco
Model Architecture: SSD FPN object detection
Network Backbone: R-50-FPN
No of Categories: 80
Dataset Used: MS COCO
24. TFlite-object-detection-v2-ssd_resnet50_v1_fpn_1024x1024_coco
Model Architecture: SSD FPN object detection
Network Backbone: R-50-FPN
No of Categories: 80
Dataset Used: MS COCO
25. YOLOv4_CSP(Cross Stage Partial Network)
Model Name: YOLOv4
No of Categories: 80
Dataset Used: MS COCO
26. YOLOv4-large
Model Name: YOLOv4
No of Categories: 80
Dataset Used: MS COCO
27. YOLOv4-tiny
Model Name: YOLOv4
No of Categories: 80
Dataset Used: MS COCO
28. YOLOv4-large-608
Model Name: YOLOv4
Input Size: 608
No of Categories: 80
Dataset Used: MS COCO
29. YOLOv4-large-leaky
Model Name: YOLOv4
Activation Function: leaky RELU
No of Categories: 80
Dataset Used: MS COCO
30. YOLOv4-large-mish
Model Name: YOLOv4
Activation Function : Mish
No of Categories: 80
Dataset Used: MS COCO
31. YOLOv4-large-608-leaky
Model Name: YOLOv4
Input Size: 608
Activation Function : leaky RELU
No of Categories: 80
Dataset Used: MS COCO
32. YOLOv4-large-608-mish
Model Name: YOLOv4
Input Size: 608
Activation Function : Mish
No of Categories: 80
Dataset Used: MS COCO
33. MTCNN-TensorFlow
Model Name: Multi-task CNN
No of Categories: 1 (Face)
Dataset Used: WIDER FACE dataset, Celeba, LFW
34. Mediapipe-Face-Detection
Model Name: BlazeFace
No of Categories: 1 (Face)
Dataset Used: Created private geographically diverse dataset
35. mmdet-RetinaNet-R50-NASFPN-50e
Model Library: mmdet
Model Architecture : RetinaNet
Network Backbone: R-50-NASFPN
lr schd: 50e
No of Categories: 80
Dataset Used: MS COCO
36. mmdet-RetinaNet-R50-FPN-50e
Model Library: mmdet
Model Architecture : RetinaNet
Network Backbone: R-50-FPN
lr schd: 50e
No of Categories: 80
Dataset Used: MS COCO
37. mmdet-RetinaNet-R101-FPN-2x
Model Library: mmdet
Model Architecture : RetinaNet
Network Backbone: R-101-FPN
lr schd: 2x
No of Categories: 80
Dataset Used: MS COCO
38. mmdet-RetinaNet-R101-FPN-1x
Model Library: mmdet
Model Architecture : RetinaNet
Network Backbone: R-101-FPN
lr schd: 1x
No of Categories: 80
Dataset Used: MS COCO
39. YOLOv3_mmdetection-PyTorch
Model Library: mmdet
Model Name: YOLOv3
Network Backbone: Darknet-53
No of Categories: 80
Dataset Used: MS COCO
40. SSD-VGG16-512-mmdet-PyTorch
Model Library: mmdet
Input Size: 512
Network Backbone: VGG16
No of Categories: 80
Dataset Used: MS COCO
41. SSD-VGG16-300-mmdet-PyTorch
Model Library: mmdet
Input Size: 300
Network Backbone: VGG16
No of Categories: 80
Dataset Used: MS COCO
42. Mask-RCNN-mmdet-R50-FPN-2x
Model Library: mmdet
Network Backbone: R-50-FPN
lr schd: 2x
No of Categories: 80
Dataset Used: MS COCO
43. Mask-RCNN-mmdet-R50-FPN-1x
Model Library: mmdet
Network Backbone: R-50-FPN
lr schd: 1x
No of Categories: 80
Dataset Used: MS COCO
44. Mask-RCNN-mmdet-R101-FPN-2x
Model Library: mmdet
Network Backbone: R-101-FPN
lr schd: 2x
No of Categories: 80
Dataset Used: MS COCO
45. Mask-RCNN-mmdet-R101-FPN-1x
Model Library: mmdet
Network Backbone: R-101-FPN
lr schd: 2x
No of Categories: 80
Dataset Used: MS COCO
46. Faster-RCNN-mmdet-R101-FPN-2x
Model Library: mmdet
Network Backbone: R-101-FPN
lr schd: 2x
No of Categories: 80
Dataset Used: MS COCO
47. Faster-RCNN-mmdet-R101-FPN-1x
Model Library: mmdet
Network Backbone: R-101-FPN
lr schd: 1x
No of Categories: 80
Dataset Used: MS COCO
48. Dynamic-RCNN-mmdet-R50
Model Library: mmdet
Network Backbone: R-50
lr schd: 1x
No of Categories: 80
Dataset Used: MS COCO
49. DETR-mmdetection-PyTorch-R50
Model Library: mmdet
Network Backbone: R-50
lr schd: 150e
No of Categories: 80
Dataset Used: MS COCO
50. Cascade-Mask-RCNN-mmdet-R50-FPN-20e
Model Library: mmdet
Network Backbone: R-50-FPN
lr schd: 20e
No of Categories: 80
Dataset Used: MS COCO
51. Cascade-Mask-RCNN-mmdet-R50-FPN-1x
Model Library: mmdet
Network Backbone: R-50-FPN
lr schd: 1x
No of Categories: 80
Dataset Used: MS COCO
52. Cascade-Mask-RCNN-mmdet-R101-FPN-20e
Model Library: mmdet
Network Backbone: R-101-FPN
lr schd: 20e
No of Categories: 80
Dataset Used: MS COCO
53. Cascade-Mask-RCNN-mmdet-R101-FPN-1x
Model Library: mmdet
Network Backbone: R-101-FPN
lr schd: 1x
No of Categories: 80
Dataset Used: MS COCO
54. YOLOv5x_TTA_COCO
Model Name: YOLOv5
No of Categories: 80
Dataset Used: MS COCO
55. YOLOv5l_COCO
Model Name: YOLOv5
No of Categories: 80
Dataset Used: MS COCO
56. YOLOv5m_COCO
Model Name: YOLOv5m
No of Categories: 80
Dataset Used: MS COCO
How to create Annotations on an Object Detection Dataset using S-O-T-A Pre-Trained Models
Steps to Follow
1. log in to the tool
2. Click on the Add button.
3. The Add Dataset Form will appear.
4. Fill in all the fields such as dataset name, dataset description, category-type, and categories.
4 a. Select category-type as Rectangle
4 b. Choose Pretraining Models from the available list.
4 c. Categories according to the chosen model will appear.
5. Add data to NeuralMarker using the available options listed below:
- Google Drive Link or S3 link
- CSV File with Image URL's
- Drag and Drop
6. Click on Submit Button
7. A New Dataset will be created.
8. After New Dataset is created:
- The Dataset card will display the status "Pretrain Model running".
- The brain symbol on hover will display "AI labeling in Progress".
9. After Pretrain Model stops running over the entire dataset:
- The Dataset card will display the status "Pretrain Model Done & Ready For Annotation".
- The brain symbol on the card will display "AI Labeling Report".
10. After Pretrain Model stops running over the entire dataset, an AI labeling report will be generated:
- With a confidence score of annotation over the images in the dataset.
