※ 引述《NTUEE2CS (电机转资工)》之铭言:
: https://www.tensorflow.org/api_docs/python/tf/keras/layers/UpSampling2D
: 辣鸡TF的upsampling居然只有nearest跟 bilinear两种
: 干你娘
: 2022了 openCV早就一堆能选
: 垃圾TF还是只有两种
: 操你妈
: 又要自己造轮子= =
干我以为叫个lambda layer然后
然后改个backend里面的resize_images就能用
结果 又去改 tf.image.resize
干你娘怎么那么复杂
https://github.com/tensorflow/tensorflow/blob/v2.8.0/tensorflow/python/ops/image_ops_impl.py#L1545-L1721
难道我得重编TF?