Commit 2d651fe

committed

add yolo docs

1 parent 6379d01 commit 2d651feCopy full SHA for 2d651fe

File tree

+51

-1

lines changed

tensorlayer/app
- computer_vision.py
- computer_vision_object_detection
  - yolov4.py

+51

-1

lines changed

Lines changed: 27 additions & 0 deletions

Original file line number	Diff line number	Diff line change
`@@ -9,6 +9,33 @@`
`9`	`9`
`10`	`10`
`11`	`11`	`class object_detection(object):`
	`12`	`+ """Model encapsulation.`
	`13`	`+`
	`14`	`+ Parameters`
	`15`	`+ ----------`
	`16`	`+ model_name : str`
	`17`	`+ Choose the model to inference.`
	`18`	`+`
	`19`	`+ Methods`
	`20`	`+ ---------`
	`21`	`+ __init__()`
	`22`	`+ Initializing the model.`
	`23`	`+ __call__()`
	`24`	`+ (1)Formatted input and output. (2)Inference model.`
	`25`	`+ list()`
	`26`	`+ Abstract method. Return available a list of model_name.`
	`27`	`+`
	`28`	`+ Examples`
	`29`	`+ ---------`
	`30`	+ Object Detection detection MSCOCO with YOLOv4, see `tutorial_object_detection_yolov4.py
	`31`	+ <https://github.com/tensorlayer/tensorlayer/blob/master/example/app_tutorials/tutorial_object_detection_yolov4.py>`__
	`32`	`+ With TensorLayer`
	`33`	`+`
	`34`	`+ >>> # get the whole model`
	`35`	`+ >>> net = tl.app.computer_vision.object_detection('yolo4-mscoco')`
	`36`	`+ >>> # use for inferencing`
	`37`	`+ >>> output = net(img)`
	`38`	`+ """`
`12`	`39`
`13`	`40`	`def __init__(self, model_name='yolo4-mscoco'):`
`14`	`41`	`self.model_name = model_name`

Lines changed: 24 additions & 1 deletion

Original file line number	Diff line number	Diff line change
`@@ -1,6 +1,6 @@`
`1`	`1`	`#! /usr/bin/python`
`2`	`2`	`# -- coding: utf-8 --`
`3`		`-"""YOLOv4 for MSCOCO.`
	`3`	`+"""YOLOv4 for MS-COCO.`
`4`	`4`
`5`	`5`	`# Reference:`
`6`	`6`	`- [tensorflow-yolov4-tflite](`
`@@ -139,6 +139,29 @@ def cspdarknet53(input_data=None):`
`139`	`139`
`140`	`140`
`141`	`141`	`def YOLOv4(NUM_CLASS, pretrained=False):`
	`142`	`+ """Pre-trained YOLOv4 model.`
	`143`	`+`
	`144`	`+ Parameters`
	`145`	`+ ------------`
	`146`	`+ NUM_CLASS : int`
	`147`	`+ Number of classes in final prediction.`
	`148`	`+ pretrained : boolean`
	`149`	`+ Whether to load pretrained weights. Default False.`
	`150`	`+`
	`151`	`+ Examples`
	`152`	`+ ---------`
	`153`	+ Object Detection with YOLOv4, see `computer_vision.py
	`154`	+ <https://github.com/tensorlayer/tensorlayer/blob/master/tensorlayer/app/computer_vision.py>`__
	`155`	`+ With TensorLayer`
	`156`	`+`
	`157`	`+ >>> # get the whole model, without pre-trained YOLOv4 parameters`
	`158`	`+ >>> yolov4 = tl.app.YOLOv4(NUM_CLASS=80, pretrained=False)`
	`159`	`+ >>> # get the whole model, restore pre-trained YOLOv4 parameters`
	`160`	`+ >>> yolov4 = tl.app.YOLOv4(NUM_CLASS=80, pretrained=True)`
	`161`	`+ >>> # use for inferencing`
	`162`	`+ >>> output = yolov4(img, is_train=False)`
	`163`	`+`
	`164`	`+ """`
`142`	`165`
`143`	`166`	`input_layer = Input([None, INPUT_SIZE, INPUT_SIZE, 3])`
`144`	`167`	`route_1, route_2, conv = cspdarknet53(input_layer)`

Comments

(0)