CV之YOLOv3：基于Tensorflow框架利用YOLOv3算法对热播新剧《庆余年》实现目标检测

CV之YOLOv3：基于Tensorflow框架利用YOLOv3算法对热播新剧《庆余年》实现目标检测

搭建

1、下载代码

tensorflow-yolov3

2、安装依赖库

pip install -r ./docs/requirements.txt

3、导出COCO权重解压到checkpoint文件夹内

Exporting loaded COCO weights as TF checkpoint(yolov3_coco.ckpt

python convert_weight.py
python freeze_graph.py

4、测试

2019-12-25 15:05:02.766745: I
=> yolov3/darknet-53/Conv/weights                     (3, 3, 3, 32)
=> yolov3/darknet-53/Conv/BatchNorm/gamma             (32,)
=> yolov3/darknet-53/Conv/BatchNorm/beta              (32,)
=> yolov3/darknet-53/Conv/BatchNorm/moving_mean       (32,)
=> yolov3/darknet-53/Conv/BatchNorm/moving_variance   (32,)
=> yolov3/darknet-53/Conv_1/weights                   (3, 3, 32, 64)
=> yolov3/darknet-53/Conv_1/BatchNorm/gamma           (64,)
=> yolov3/darknet-53/Conv_1/BatchNorm/beta            (64,)
=> yolov3/darknet-53/Conv_1/BatchNorm/moving_mean     (64,)
=> yolov3/darknet-53/Conv_1/BatchNorm/moving_variance (64,)
=> yolov3/darknet-53/Conv_2/weights                   (1, 1, 64, 32)
=> yolov3/darknet-53/Conv_2/BatchNorm/gamma           (32,)
=> yolov3/darknet-53/Conv_2/BatchNorm/beta            (32,)
=> yolov3/darknet-53/Conv_2/BatchNorm/moving_mean     (32,)
=> yolov3/darknet-53/Conv_2/BatchNorm/moving_variance (32,)
=> yolov3/darknet-53/Conv_3/weights                   (3, 3, 32, 64)
=> yolov3/darknet-53/Conv_3/BatchNorm/gamma           (64,)
=> yolov3/darknet-53/Conv_3/BatchNorm/beta            (64,)
=> yolov3/darknet-53/Conv_3/BatchNorm/moving_mean     (64,)
=> yolov3/darknet-53/Conv_3/BatchNorm/moving_variance (64,)
=> yolov3/darknet-53/Conv_4/weights                   (3, 3, 64, 128)
=> yolov3/darknet-53/Conv_4/BatchNorm/gamma           (128,)
=> yolov3/darknet-53/Conv_4/BatchNorm/beta            (128,)
=> yolov3/darknet-53/Conv_4/BatchNorm/moving_mean     (128,)
=> yolov3/darknet-53/Conv_4/BatchNorm/moving_variance (128,)
=> yolov3/darknet-53/Conv_5/weights                   (1, 1, 128, 64)
=> yolov3/darknet-53/Conv_5/BatchNorm/gamma           (64,)
=> yolov3/darknet-53/Conv_5/BatchNorm/beta            (64,)
=> yolov3/darknet-53/Conv_5/BatchNorm/moving_mean     (64,)
=> yolov3/darknet-53/Conv_5/BatchNorm/moving_variance (64,)
=> yolov3/darknet-53/Conv_6/weights                   (3, 3, 64, 128)
=> yolov3/darknet-53/Conv_6/BatchNorm/gamma           (128,)
=> yolov3/darknet-53/Conv_6/BatchNorm/beta            (128,)
=> yolov3/darknet-53/Conv_6/BatchNorm/moving_mean     (128,)
=> yolov3/darknet-53/Conv_6/BatchNorm/moving_variance (128,)
=> yolov3/darknet-53/Conv_7/weights                   (1, 1, 128, 64)
=> yolov3/darknet-53/Conv_7/BatchNorm/gamma           (64,)
=> yolov3/darknet-53/Conv_7/BatchNorm/beta            (64,)
=> yolov3/darknet-53/Conv_7/BatchNorm/moving_mean     (64,)
=> yolov3/darknet-53/Conv_7/BatchNorm/moving_variance (64,)
=> yolov3/darknet-53/Conv_8/weights                   (3, 3, 64, 128)
=> yolov3/darknet-53/Conv_8/BatchNorm/gamma           (128,)
=> yolov3/darknet-53/Conv_8/BatchNorm/beta            (128,)
=> yolov3/darknet-53/Conv_8/BatchNorm/moving_mean     (128,)
=> yolov3/darknet-53/Conv_8/BatchNorm/moving_variance (128,)
=> yolov3/darknet-53/Conv_9/weights                   (3, 3, 128, 256)
=> yolov3/darknet-53/Conv_9/BatchNorm/gamma           (256,)
=> yolov3/darknet-53/Conv_9/BatchNorm/beta            (256,)
=> yolov3/darknet-53/Conv_9/BatchNorm/moving_mean     (256,)
=> yolov3/darknet-53/Conv_9/BatchNorm/moving_variance (256,)
=> yolov3/darknet-53/Conv_10/weights                  (1, 1, 256, 128)
=> yolov3/darknet-53/Conv_10/BatchNorm/gamma          (128,)
=> yolov3/darknet-53/Conv_10/BatchNorm/beta           (128,)
=> yolov3/darknet-53/Conv_10/BatchNorm/moving_mean    (128,)
=> yolov3/darknet-53/Conv_10/BatchNorm/moving_variance (128,)
=> yolov3/darknet-53/Conv_11/weights                  (3, 3, 128, 256)
=> yolov3/darknet-53/Conv_11/BatchNorm/gamma          (256,)
=> yolov3/darknet-53/Conv_11/BatchNorm/beta           (256,)
=> yolov3/darknet-53/Conv_11/BatchNorm/moving_mean    (256,)
=> yolov3/darknet-53/Conv_11/BatchNorm/moving_variance (256,)
=> yolov3/darknet-53/Conv_12/weights                  (1, 1, 256, 128)
=> yolov3/darknet-53/Conv_12/BatchNorm/gamma          (128,)
=> yolov3/darknet-53/Conv_12/BatchNorm/beta           (128,)
=> yolov3/darknet-53/Conv_12/BatchNorm/moving_mean    (128,)
=> yolov3/darknet-53/Conv_12/BatchNorm/moving_variance (128,)
=> yolov3/darknet-53/Conv_13/weights                  (3, 3, 128, 256)
=> yolov3/darknet-53/Conv_13/BatchNorm/gamma          (256,)
=> yolov3/darknet-53/Conv_13/BatchNorm/beta           (256,)
=> yolov3/darknet-53/Conv_13/BatchNorm/moving_mean    (256,)
=> yolov3/darknet-53/Conv_13/BatchNorm/moving_variance (256,)
=> yolov3/darknet-53/Conv_14/weights                  (1, 1, 256, 128)
=> yolov3/darknet-53/Conv_14/BatchNorm/gamma          (128,)
=> yolov3/darknet-53/Conv_14/BatchNorm/beta           (128,)
=> yolov3/darknet-53/Conv_14/BatchNorm/moving_mean    (128,)
=> yolov3/darknet-53/Conv_14/BatchNorm/moving_variance (128,)
=> yolov3/darknet-53/Conv_15/weights                  (3, 3, 128, 256)
=> yolov3/darknet-53/Conv_15/BatchNorm/gamma          (256,)
=> yolov3/darknet-53/Conv_15/BatchNorm/beta           (256,)
=> yolov3/darknet-53/Conv_15/BatchNorm/moving_mean    (256,)
=> yolov3/darknet-53/Conv_15/BatchNorm/moving_variance (256,)
=> yolov3/darknet-53/Conv_16/weights                  (1, 1, 256, 128)
=> yolov3/darknet-53/Conv_16/BatchNorm/gamma          (128,)
=> yolov3/darknet-53/Conv_16/BatchNorm/beta           (128,)
=> yolov3/darknet-53/Conv_16/BatchNorm/moving_mean    (128,)
=> yolov3/darknet-53/Conv_16/BatchNorm/moving_variance (128,)
=> yolov3/darknet-53/Conv_17/weights                  (3, 3, 128, 256)
=> yolov3/darknet-53/Conv_17/BatchNorm/gamma          (256,)
=> yolov3/darknet-53/Conv_17/BatchNorm/beta           (256,)
=> yolov3/darknet-53/Conv_17/BatchNorm/moving_mean    (256,)
=> yolov3/darknet-53/Conv_17/BatchNorm/moving_variance (256,)
=> yolov3/darknet-53/Conv_18/weights                  (1, 1, 256, 128)
=> yolov3/darknet-53/Conv_18/BatchNorm/gamma          (128,)
=> yolov3/darknet-53/Conv_18/BatchNorm/beta           (128,)
=> yolov3/darknet-53/Conv_18/BatchNorm/moving_mean    (128,)
=> yolov3/darknet-53/Conv_18/BatchNorm/moving_variance (128,)
=> yolov3/darknet-53/Conv_19/weights                  (3, 3, 128, 256)
=> yolov3/darknet-53/Conv_19/BatchNorm/gamma          (256,)
=> yolov3/darknet-53/Conv_19/BatchNorm/beta           (256,)
=> yolov3/darknet-53/Conv_19/BatchNorm/moving_mean    (256,)
=> yolov3/darknet-53/Conv_19/BatchNorm/moving_variance (256,)
=> yolov3/darknet-53/Conv_20/weights                  (1, 1, 256, 128)
=> yolov3/darknet-53/Conv_20/BatchNorm/gamma          (128,)
=> yolov3/darknet-53/Conv_20/BatchNorm/beta           (128,)
=> yolov3/darknet-53/Conv_20/BatchNorm/moving_mean    (128,)
=> yolov3/darknet-53/Conv_20/BatchNorm/moving_variance (128,)
=> yolov3/darknet-53/Conv_21/weights                  (3, 3, 128, 256)
=> yolov3/darknet-53/Conv_21/BatchNorm/gamma          (256,)
=> yolov3/darknet-53/Conv_21/BatchNorm/beta           (256,)
=> yolov3/darknet-53/Conv_21/BatchNorm/moving_mean    (256,)
=> yolov3/darknet-53/Conv_21/BatchNorm/moving_variance (256,)
=> yolov3/darknet-53/Conv_22/weights                  (1, 1, 256, 128)
=> yolov3/darknet-53/Conv_22/BatchNorm/gamma          (128,)
=> yolov3/darknet-53/Conv_22/BatchNorm/beta           (128,)
=> yolov3/darknet-53/Conv_22/BatchNorm/moving_mean    (128,)
=> yolov3/darknet-53/Conv_22/BatchNorm/moving_variance (128,)
=> yolov3/darknet-53/Conv_23/weights                  (3, 3, 128, 256)
=> yolov3/darknet-53/Conv_23/BatchNorm/gamma          (256,)
=> yolov3/darknet-53/Conv_23/BatchNorm/beta           (256,)
=> yolov3/darknet-53/Conv_23/BatchNorm/moving_mean    (256,)
=> yolov3/darknet-53/Conv_23/BatchNorm/moving_variance (256,)
=> yolov3/darknet-53/Conv_24/weights                  (1, 1, 256, 128)
=> yolov3/darknet-53/Conv_24/BatchNorm/gamma          (128,)
=> yolov3/darknet-53/Conv_24/BatchNorm/beta           (128,)
=> yolov3/darknet-53/Conv_24/BatchNorm/moving_mean    (128,)
=> yolov3/darknet-53/Conv_24/BatchNorm/moving_variance (128,)
=> yolov3/darknet-53/Conv_25/weights                  (3, 3, 128, 256)
=> yolov3/darknet-53/Conv_25/BatchNorm/gamma          (256,)
=> yolov3/darknet-53/Conv_25/BatchNorm/beta           (256,)
=> yolov3/darknet-53/Conv_25/BatchNorm/moving_mean    (256,)
=> yolov3/darknet-53/Conv_25/BatchNorm/moving_variance (256,)
=> yolov3/darknet-53/Conv_26/weights                  (3, 3, 256, 512)
=> yolov3/darknet-53/Conv_26/BatchNorm/gamma          (512,)
=> yolov3/darknet-53/Conv_26/BatchNorm/beta           (512,)
=> yolov3/darknet-53/Conv_26/BatchNorm/moving_mean    (512,)
=> yolov3/darknet-53/Conv_26/BatchNorm/moving_variance (512,)
=> yolov3/darknet-53/Conv_27/weights                  (1, 1, 512, 256)
=> yolov3/darknet-53/Conv_27/BatchNorm/gamma          (256,)
=> yolov3/darknet-53/Conv_27/BatchNorm/beta           (256,)
=> yolov3/darknet-53/Conv_27/BatchNorm/moving_mean    (256,)
=> yolov3/darknet-53/Conv_27/BatchNorm/moving_variance (256,)
=> yolov3/darknet-53/Conv_28/weights                  (3, 3, 256, 512)
=> yolov3/darknet-53/Conv_28/BatchNorm/gamma          (512,)
=> yolov3/darknet-53/Conv_28/BatchNorm/beta           (512,)
=> yolov3/darknet-53/Conv_28/BatchNorm/moving_mean    (512,)
=> yolov3/darknet-53/Conv_28/BatchNorm/moving_variance (512,)
=> yolov3/darknet-53/Conv_29/weights                  (1, 1, 512, 256)
=> yolov3/darknet-53/Conv_29/BatchNorm/gamma          (256,)
=> yolov3/darknet-53/Conv_29/BatchNorm/beta           (256,)
=> yolov3/darknet-53/Conv_29/BatchNorm/moving_mean    (256,)
=> yolov3/darknet-53/Conv_29/BatchNorm/moving_variance (256,)
=> yolov3/darknet-53/Conv_30/weights                  (3, 3, 256, 512)
=> yolov3/darknet-53/Conv_30/BatchNorm/gamma          (512,)
=> yolov3/darknet-53/Conv_30/BatchNorm/beta           (512,)
=> yolov3/darknet-53/Conv_30/BatchNorm/moving_mean    (512,)
=> yolov3/darknet-53/Conv_30/BatchNorm/moving_variance (512,)
=> yolov3/darknet-53/Conv_31/weights                  (1, 1, 512, 256)
=> yolov3/darknet-53/Conv_31/BatchNorm/gamma          (256,)
=> yolov3/darknet-53/Conv_31/BatchNorm/beta           (256,)
=> yolov3/darknet-53/Conv_31/BatchNorm/moving_mean    (256,)
=> yolov3/darknet-53/Conv_31/BatchNorm/moving_variance (256,)
=> yolov3/darknet-53/Conv_32/weights                  (3, 3, 256, 512)
=> yolov3/darknet-53/Conv_32/BatchNorm/gamma          (512,)
=> yolov3/darknet-53/Conv_32/BatchNorm/beta           (512,)
=> yolov3/darknet-53/Conv_32/BatchNorm/moving_mean    (512,)
=> yolov3/darknet-53/Conv_32/BatchNorm/moving_variance (512,)
=> yolov3/darknet-53/Conv_33/weights                  (1, 1, 512, 256)
=> yolov3/darknet-53/Conv_33/BatchNorm/gamma          (256,)
=> yolov3/darknet-53/Conv_33/BatchNorm/beta           (256,)
=> yolov3/darknet-53/Conv_33/BatchNorm/moving_mean    (256,)
=> yolov3/darknet-53/Conv_33/BatchNorm/moving_variance (256,)
=> yolov3/darknet-53/Conv_34/weights                  (3, 3, 256, 512)
=> yolov3/darknet-53/Conv_34/BatchNorm/gamma          (512,)
=> yolov3/darknet-53/Conv_34/BatchNorm/beta           (512,)
=> yolov3/darknet-53/Conv_34/BatchNorm/moving_mean    (512,)
=> yolov3/darknet-53/Conv_34/BatchNorm/moving_variance (512,)
=> yolov3/darknet-53/Conv_35/weights                  (1, 1, 512, 256)
=> yolov3/darknet-53/Conv_35/BatchNorm/gamma          (256,)
=> yolov3/darknet-53/Conv_35/BatchNorm/beta           (256,)
=> yolov3/darknet-53/Conv_35/BatchNorm/moving_mean    (256,)
=> yolov3/darknet-53/Conv_35/BatchNorm/moving_variance (256,)
=> yolov3/darknet-53/Conv_36/weights                  (3, 3, 256, 512)
=> yolov3/darknet-53/Conv_36/BatchNorm/gamma          (512,)
=> yolov3/darknet-53/Conv_36/BatchNorm/beta           (512,)
=> yolov3/darknet-53/Conv_36/BatchNorm/moving_mean    (512,)
=> yolov3/darknet-53/Conv_36/BatchNorm/moving_variance (512,)
=> yolov3/darknet-53/Conv_37/weights                  (1, 1, 512, 256)
=> yolov3/darknet-53/Conv_37/BatchNorm/gamma          (256,)
=> yolov3/darknet-53/Conv_37/BatchNorm/beta           (256,)
=> yolov3/darknet-53/Conv_37/BatchNorm/moving_mean    (256,)
=> yolov3/darknet-53/Conv_37/BatchNorm/moving_variance (256,)
=> yolov3/darknet-53/Conv_38/weights                  (3, 3, 256, 512)
=> yolov3/darknet-53/Conv_38/BatchNorm/gamma          (512,)
=> yolov3/darknet-53/Conv_38/BatchNorm/beta           (512,)
=> yolov3/darknet-53/Conv_38/BatchNorm/moving_mean    (512,)
=> yolov3/darknet-53/Conv_38/BatchNorm/moving_variance (512,)
=> yolov3/darknet-53/Conv_39/weights                  (1, 1, 512, 256)
=> yolov3/darknet-53/Conv_39/BatchNorm/gamma          (256,)
=> yolov3/darknet-53/Conv_39/BatchNorm/beta           (256,)
=> yolov3/darknet-53/Conv_39/BatchNorm/moving_mean    (256,)
=> yolov3/darknet-53/Conv_39/BatchNorm/moving_variance (256,)
=> yolov3/darknet-53/Conv_40/weights                  (3, 3, 256, 512)
=> yolov3/darknet-53/Conv_40/BatchNorm/gamma          (512,)
=> yolov3/darknet-53/Conv_40/BatchNorm/beta           (512,)
=> yolov3/darknet-53/Conv_40/BatchNorm/moving_mean    (512,)
=> yolov3/darknet-53/Conv_40/BatchNorm/moving_variance (512,)
=> yolov3/darknet-53/Conv_41/weights                  (1, 1, 512, 256)
=> yolov3/darknet-53/Conv_41/BatchNorm/gamma          (256,)
=> yolov3/darknet-53/Conv_41/BatchNorm/beta           (256,)
=> yolov3/darknet-53/Conv_41/BatchNorm/moving_mean    (256,)
=> yolov3/darknet-53/Conv_41/BatchNorm/moving_variance (256,)
=> yolov3/darknet-53/Conv_42/weights                  (3, 3, 256, 512)
=> yolov3/darknet-53/Conv_42/BatchNorm/gamma          (512,)
=> yolov3/darknet-53/Conv_42/BatchNorm/beta           (512,)
=> yolov3/darknet-53/Conv_42/BatchNorm/moving_mean    (512,)
=> yolov3/darknet-53/Conv_42/BatchNorm/moving_variance (512,)
=> yolov3/darknet-53/Conv_43/weights                  (3, 3, 512, 1024)
=> yolov3/darknet-53/Conv_43/BatchNorm/gamma          (1024,)
=> yolov3/darknet-53/Conv_43/BatchNorm/beta           (1024,)
=> yolov3/darknet-53/Conv_43/BatchNorm/moving_mean    (1024,)
=> yolov3/darknet-53/Conv_43/BatchNorm/moving_variance (1024,)
=> yolov3/darknet-53/Conv_44/weights                  (1, 1, 1024, 512)
=> yolov3/darknet-53/Conv_44/BatchNorm/gamma          (512,)
=> yolov3/darknet-53/Conv_44/BatchNorm/beta           (512,)
=> yolov3/darknet-53/Conv_44/BatchNorm/moving_mean    (512,)
=> yolov3/darknet-53/Conv_44/BatchNorm/moving_variance (512,)
=> yolov3/darknet-53/Conv_45/weights                  (3, 3, 512, 1024)
=> yolov3/darknet-53/Conv_45/BatchNorm/gamma          (1024,)
=> yolov3/darknet-53/Conv_45/BatchNorm/beta           (1024,)
=> yolov3/darknet-53/Conv_45/BatchNorm/moving_mean    (1024,)
=> yolov3/darknet-53/Conv_45/BatchNorm/moving_variance (1024,)
=> yolov3/darknet-53/Conv_46/weights                  (1, 1, 1024, 512)
=> yolov3/darknet-53/Conv_46/BatchNorm/gamma          (512,)
=> yolov3/darknet-53/Conv_46/BatchNorm/beta           (512,)
=> yolov3/darknet-53/Conv_46/BatchNorm/moving_mean    (512,)
=> yolov3/darknet-53/Conv_46/BatchNorm/moving_variance (512,)
=> yolov3/darknet-53/Conv_47/weights                  (3, 3, 512, 1024)
=> yolov3/darknet-53/Conv_47/BatchNorm/gamma          (1024,)
=> yolov3/darknet-53/Conv_47/BatchNorm/beta           (1024,)
=> yolov3/darknet-53/Conv_47/BatchNorm/moving_mean    (1024,)
=> yolov3/darknet-53/Conv_47/BatchNorm/moving_variance (1024,)
=> yolov3/darknet-53/Conv_48/weights                  (1, 1, 1024, 512)
=> yolov3/darknet-53/Conv_48/BatchNorm/gamma          (512,)
=> yolov3/darknet-53/Conv_48/BatchNorm/beta           (512,)
=> yolov3/darknet-53/Conv_48/BatchNorm/moving_mean    (512,)
=> yolov3/darknet-53/Conv_48/BatchNorm/moving_variance (512,)
=> yolov3/darknet-53/Conv_49/weights                  (3, 3, 512, 1024)
=> yolov3/darknet-53/Conv_49/BatchNorm/gamma          (1024,)
=> yolov3/darknet-53/Conv_49/BatchNorm/beta           (1024,)
=> yolov3/darknet-53/Conv_49/BatchNorm/moving_mean    (1024,)
=> yolov3/darknet-53/Conv_49/BatchNorm/moving_variance (1024,)
=> yolov3/darknet-53/Conv_50/weights                  (1, 1, 1024, 512)
=> yolov3/darknet-53/Conv_50/BatchNorm/gamma          (512,)
=> yolov3/darknet-53/Conv_50/BatchNorm/beta           (512,)
=> yolov3/darknet-53/Conv_50/BatchNorm/moving_mean    (512,)
=> yolov3/darknet-53/Conv_50/BatchNorm/moving_variance (512,)
=> yolov3/darknet-53/Conv_51/weights                  (3, 3, 512, 1024)
=> yolov3/darknet-53/Conv_51/BatchNorm/gamma          (1024,)
=> yolov3/darknet-53/Conv_51/BatchNorm/beta           (1024,)
=> yolov3/darknet-53/Conv_51/BatchNorm/moving_mean    (1024,)
=> yolov3/darknet-53/Conv_51/BatchNorm/moving_variance (1024,)
=> yolov3/yolo-v3/Conv/weights                        (1, 1, 1024, 512)
=> yolov3/yolo-v3/Conv/BatchNorm/gamma                (512,)
=> yolov3/yolo-v3/Conv/BatchNorm/beta                 (512,)
=> yolov3/yolo-v3/Conv/BatchNorm/moving_mean          (512,)
=> yolov3/yolo-v3/Conv/BatchNorm/moving_variance      (512,)
=> yolov3/yolo-v3/Conv_1/weights                      (3, 3, 512, 1024)
=> yolov3/yolo-v3/Conv_1/BatchNorm/gamma              (1024,)
=> yolov3/yolo-v3/Conv_1/BatchNorm/beta               (1024,)
=> yolov3/yolo-v3/Conv_1/BatchNorm/moving_mean        (1024,)
=> yolov3/yolo-v3/Conv_1/BatchNorm/moving_variance    (1024,)
=> yolov3/yolo-v3/Conv_2/weights                      (1, 1, 1024, 512)
=> yolov3/yolo-v3/Conv_2/BatchNorm/gamma              (512,)
=> yolov3/yolo-v3/Conv_2/BatchNorm/beta               (512,)
=> yolov3/yolo-v3/Conv_2/BatchNorm/moving_mean        (512,)
=> yolov3/yolo-v3/Conv_2/BatchNorm/moving_variance    (512,)
=> yolov3/yolo-v3/Conv_3/weights                      (3, 3, 512, 1024)
=> yolov3/yolo-v3/Conv_3/BatchNorm/gamma              (1024,)
=> yolov3/yolo-v3/Conv_3/BatchNorm/beta               (1024,)
=> yolov3/yolo-v3/Conv_3/BatchNorm/moving_mean        (1024,)
=> yolov3/yolo-v3/Conv_3/BatchNorm/moving_variance    (1024,)
=> yolov3/yolo-v3/Conv_4/weights                      (1, 1, 1024, 512)
=> yolov3/yolo-v3/Conv_4/BatchNorm/gamma              (512,)
=> yolov3/yolo-v3/Conv_4/BatchNorm/beta               (512,)
=> yolov3/yolo-v3/Conv_4/BatchNorm/moving_mean        (512,)
=> yolov3/yolo-v3/Conv_4/BatchNorm/moving_variance    (512,)
=> yolov3/yolo-v3/Conv_5/weights                      (3, 3, 512, 1024)
=> yolov3/yolo-v3/Conv_5/BatchNorm/gamma              (1024,)
=> yolov3/yolo-v3/Conv_5/BatchNorm/beta               (1024,)
=> yolov3/yolo-v3/Conv_5/BatchNorm/moving_mean        (1024,)
=> yolov3/yolo-v3/Conv_5/BatchNorm/moving_variance    (1024,)
=> yolov3/yolo-v3/Conv_6/weights                      (1, 1, 1024, 255)
=> yolov3/yolo-v3/Conv_6/biases                       (255,)
=> yolov3/yolo-v3/Conv_7/weights                      (1, 1, 512, 256)
=> yolov3/yolo-v3/Conv_7/BatchNorm/gamma              (256,)
=> yolov3/yolo-v3/Conv_7/BatchNorm/beta               (256,)
=> yolov3/yolo-v3/Conv_7/BatchNorm/moving_mean        (256,)
=> yolov3/yolo-v3/Conv_7/BatchNorm/moving_variance    (256,)
=> yolov3/yolo-v3/Conv_8/weights                      (1, 1, 768, 256)
=> yolov3/yolo-v3/Conv_8/BatchNorm/gamma              (256,)
=> yolov3/yolo-v3/Conv_8/BatchNorm/beta               (256,)
=> yolov3/yolo-v3/Conv_8/BatchNorm/moving_mean        (256,)
=> yolov3/yolo-v3/Conv_8/BatchNorm/moving_variance    (256,)
=> yolov3/yolo-v3/Conv_9/weights                      (3, 3, 256, 512)
=> yolov3/yolo-v3/Conv_9/BatchNorm/gamma              (512,)
=> yolov3/yolo-v3/Conv_9/BatchNorm/beta               (512,)
=> yolov3/yolo-v3/Conv_9/BatchNorm/moving_mean        (512,)
=> yolov3/yolo-v3/Conv_9/BatchNorm/moving_variance    (512,)
=> yolov3/yolo-v3/Conv_10/weights                     (1, 1, 512, 256)
=> yolov3/yolo-v3/Conv_10/BatchNorm/gamma             (256,)
=> yolov3/yolo-v3/Conv_10/BatchNorm/beta              (256,)
=> yolov3/yolo-v3/Conv_10/BatchNorm/moving_mean       (256,)
=> yolov3/yolo-v3/Conv_10/BatchNorm/moving_variance   (256,)
=> yolov3/yolo-v3/Conv_11/weights                     (3, 3, 256, 512)
=> yolov3/yolo-v3/Conv_11/BatchNorm/gamma             (512,)
=> yolov3/yolo-v3/Conv_11/BatchNorm/beta              (512,)
=> yolov3/yolo-v3/Conv_11/BatchNorm/moving_mean       (512,)
=> yolov3/yolo-v3/Conv_11/BatchNorm/moving_variance   (512,)
=> yolov3/yolo-v3/Conv_12/weights                     (1, 1, 512, 256)
=> yolov3/yolo-v3/Conv_12/BatchNorm/gamma             (256,)
=> yolov3/yolo-v3/Conv_12/BatchNorm/beta              (256,)
=> yolov3/yolo-v3/Conv_12/BatchNorm/moving_mean       (256,)
=> yolov3/yolo-v3/Conv_12/BatchNorm/moving_variance   (256,)
=> yolov3/yolo-v3/Conv_13/weights                     (3, 3, 256, 512)
=> yolov3/yolo-v3/Conv_13/BatchNorm/gamma             (512,)
=> yolov3/yolo-v3/Conv_13/BatchNorm/beta              (512,)
=> yolov3/yolo-v3/Conv_13/BatchNorm/moving_mean       (512,)
=> yolov3/yolo-v3/Conv_13/BatchNorm/moving_variance   (512,)
=> yolov3/yolo-v3/Conv_14/weights                     (1, 1, 512, 255)
=> yolov3/yolo-v3/Conv_14/biases                      (255,)
=> yolov3/yolo-v3/Conv_15/weights                     (1, 1, 256, 128)
=> yolov3/yolo-v3/Conv_15/BatchNorm/gamma             (128,)
=> yolov3/yolo-v3/Conv_15/BatchNorm/beta              (128,)
=> yolov3/yolo-v3/Conv_15/BatchNorm/moving_mean       (128,)
=> yolov3/yolo-v3/Conv_15/BatchNorm/moving_variance   (128,)
=> yolov3/yolo-v3/Conv_16/weights                     (1, 1, 384, 128)
=> yolov3/yolo-v3/Conv_16/BatchNorm/gamma             (128,)
=> yolov3/yolo-v3/Conv_16/BatchNorm/beta              (128,)
=> yolov3/yolo-v3/Conv_16/BatchNorm/moving_mean       (128,)
=> yolov3/yolo-v3/Conv_16/BatchNorm/moving_variance   (128,)
=> yolov3/yolo-v3/Conv_17/weights                     (3, 3, 128, 256)
=> yolov3/yolo-v3/Conv_17/BatchNorm/gamma             (256,)
=> yolov3/yolo-v3/Conv_17/BatchNorm/beta              (256,)
=> yolov3/yolo-v3/Conv_17/BatchNorm/moving_mean       (256,)
=> yolov3/yolo-v3/Conv_17/BatchNorm/moving_variance   (256,)
=> yolov3/yolo-v3/Conv_18/weights                     (1, 1, 256, 128)
=> yolov3/yolo-v3/Conv_18/BatchNorm/gamma             (128,)
=> yolov3/yolo-v3/Conv_18/BatchNorm/beta              (128,)
=> yolov3/yolo-v3/Conv_18/BatchNorm/moving_mean       (128,)
=> yolov3/yolo-v3/Conv_18/BatchNorm/moving_variance   (128,)
=> yolov3/yolo-v3/Conv_19/weights                     (3, 3, 128, 256)
=> yolov3/yolo-v3/Conv_19/BatchNorm/gamma             (256,)
=> yolov3/yolo-v3/Conv_19/BatchNorm/beta              (256,)
=> yolov3/yolo-v3/Conv_19/BatchNorm/moving_mean       (256,)
=> yolov3/yolo-v3/Conv_19/BatchNorm/moving_variance   (256,)
=> yolov3/yolo-v3/Conv_20/weights                     (1, 1, 256, 128)
=> yolov3/yolo-v3/Conv_20/BatchNorm/gamma             (128,)
=> yolov3/yolo-v3/Conv_20/BatchNorm/beta              (128,)
=> yolov3/yolo-v3/Conv_20/BatchNorm/moving_mean       (128,)
=> yolov3/yolo-v3/Conv_20/BatchNorm/moving_variance   (128,)
=> yolov3/yolo-v3/Conv_21/weights                     (3, 3, 128, 256)
=> yolov3/yolo-v3/Conv_21/BatchNorm/gamma             (256,)
=> yolov3/yolo-v3/Conv_21/BatchNorm/beta              (256,)
=> yolov3/yolo-v3/Conv_21/BatchNorm/moving_mean       (256,)
=> yolov3/yolo-v3/Conv_21/BatchNorm/moving_variance   (256,)
=> yolov3/yolo-v3/Conv_22/weights                     (1, 1, 256, 255)
=> yolov3/yolo-v3/Conv_22/biases                      (255,)


Tensor("conv_sbbox/BiasAdd:0", shape=(?, ?, ?, 255), dtype=float32) Tensor("conv_mbbox/BiasAdd:0", shape=(?, ?, ?, 255), dtype=float32) Tensor("conv_lbbox/BiasAdd:0", shape=(?, ?, ?, 255), dtype=float32)
赞 (0)

相关推荐

  • c 调用yolov4模型进行目标检测

    前言 yolo系列用c写的,在工程中的部署特别方便.4月份yolov4横空出世,之前试了试效果,精度确实有了很大的提升,AB大神nb.最近需要在C++项目中使用yolov4,尝试了opencv的调用( ...

  • PP-YOLO何许模型？竟然超越了YOLOv4

    重磅干货,第一时间送达 PP-YOLO评估显示出更快的推断(x轴)和更好的准确性(y轴) PP-YOLO评估指标显示出比现有的最新对象检测模型YOLOv4更高的性能.但是,提出者百度却谦虚的声明: 无 ...

  • (4条消息) OpenCV DNN之YOLO实时对象检测

    OpenCV在3.3.1的版本中开始正式支持Darknet网络框架并且支持YOLO1与YOLO2以及YOLO Tiny网络模型的导入与使用.YOLO是一种比SSD还要快的对象检测网络模型,算法作者在其 ...

  • 【目标检测代码实战】从零开始动手实现yolov3：训练篇（一）

    前言 在前面几篇文章中小糖豆为大家讲解了yolo系列算法的演变.俗话说,光说不练假把式.接下来小糖豆将带领大家从零开始,亲自动手实现yolov3的训练与预测. 本教程说明: 需要读者已经基本了解pyt ...

  • yolov4——训练自己的数据集(完整版)

    好了,多的也不说,直接开始吧! 1.下载预训练的好权重文件 yolov4.conv.137,放在build/darknet/x64/下 2.创建配置文件:在darknet-master/cfg/下,创 ...

  • 吊打一切现有版本的YOLO！旷视重磅开源YOLOX：新一代目标检测性能速度担当！

    作者丨happy 编辑丨极市平台 极市导读 YOLO系列终于又回到了Anchor-free的怀抱,不用费劲心思去设计anchor了!旷视开源新的高性能检测器YOLOX,本文将近两年来目标检测领域的各个 ...

  • 使用Python和YOLO检测车牌

    重磅干货,第一时间送达 计算机视觉无处不在-从面部识别,制造,农业到自动驾驶汽车.今天,我们将通过动手实践进入现代计算机视觉世界,学习如何使用YOLO算法检测车牌. 来自Pexels的mali mae ...

  • 基于Opencv4.4的YOLOv4目标检测

    2020年7月18日,OpenCV官网发布了OpenCV的最新版本OpenCV4.4.0,令我比较兴奋的是,其中支持了YOLOv4,之前的一段时间,我都在YOLO系列苦苦挣扎,虽然YOLOv4的性能很 ...

  • opencv DNN模块之YOLO（Darknet）对象检测

    原理 自行百度,本人不擅长 YOLO 来自darknet对象检测框架 基于COCO数据集,能检测80个类别 YOLO V3版本 https://pireddie.com/darknet/yolo 每种 ...

  • [OpenCV实战]7 使用YOLOv3和OpenCV进行基于深度学习的目标检测

    目录 1 YOLO介绍 1.1 YOLOv3原理 1.2 为什么要将OpenCV用于YOLO? 1.3 在Darknet和OpenCV上对YOLOv3进行速度测试 2 使用YOLOv3进行对象检测(C ...

  • 使用opencv

    终于等到python版本的opencv4.4版本了,不用麻烦的编译opencv4.4,就可以调用yolov4模型了.昨天更新的版本! 在opencv4.4出来之前,通过编译darknet来推理yolo ...