calmiLovesAI/YOLOv3_TensorFlow2

Issues

• read_txt.py是有问题的

  #19 opened 4 years ago by nuaasxr
  0

• 使用COCO数据集训练，loss很快变为nan

  #23 opened 3 years ago by Guofeng-Lin
  3

• 您好,我在解码时遇到了一些问题

  #4 opened 5 years ago by Sssyz
  3

• 在运行train_from_scratch.py文件时遇到了问题

  #16 opened 4 years ago by 412869434
  1

• a small bug on test_on_single_image.py

  #21 opened 4 years ago by oujin
  1

• 图片格式为png，修改tf.io.decode_png之后还是报错

  #22 opened 4 years ago by Xiaomin-Yu
  1

• 作者你没有用上官方的预训练权重吧?

  #10 opened 5 years ago by Axe-sue
  2

• 你好，能分享一下你训练好的模型参数吗？

  #20 opened 4 years ago by Fannjh
  0

• UnicodeDecodeError: 'utf-8' codec can't decode byte 0xd5 in position 50: invalid continuation byte

  #6 opened 5 years ago by xfchangwei
  3

• 可以训练，但是使用test_on_single_image.py测试训练好的模型没有bboxs，scores,

  #2 opened 5 years ago by pursuityu
  10

• 您好，在使用您的网络时发现box_scores异常低导致 单图检测时不输出boxes

  #15 opened 4 years ago by laplace674
  1

• 训练集损失很低，测试训练集照片，框都是乱的

  #18 opened 4 years ago by YuhangJi
  0

• 请问图片训练前要resize吗？

  #13 opened 4 years ago by only-zzx
  1

• 用tf.metrics.Mean()记录损失，进入新的epoch时损失出现跳跃性下降。

  #17 opened 4 years ago by YuhangJi
  0

• 编译TFlite时，failed while converting

  #14 opened 4 years ago by only-zzx
  1

• confused with ignore_mask and true_object_mask

  #12 opened 4 years ago by speeding-motor
  1

• 求loss损失函数时，为什么用binary_crossentropy而不用softmax或sigmoid?

  #8 opened 4 years ago by Flowingsun007
  2

• 您好，发现测试时，输出没有预测边框

  #9 opened 5 years ago by deemoo-wang
  2

• how to decide the anchor box width and height

  #11 opened 5 years ago by speeding-motor
  0

• test_on_single_image.py ，仅输出原图像，无结果和框，求问，谢谢。

  #5 opened 5 years ago by peteryvqi
  1

• 单一图片测试，图片无输出，但

  #7 opened 5 years ago by TaoTiJJ
  3

• 你好，请问有把backbone换成mobilenet尝试看看效果吗？

  #3 opened 5 years ago by aa12356jm
  2

• Using a `tf.Tensor` as a Python `bool` is not allowed.

  #1 opened 5 years ago by pursuityu
  2