FPN_Tensorflow

FPN_Tensorflow copied to clipboard

如何把 FPN+ResNet 改成 FPN+MobileNet ？

2

https://github.com/DetectionTeamUCAS/FPN_Tensorflow/issues/81 我看了这个issue，还是不知道怎么改？为什么只是在cfgs.py里面写上VERSION和NET_NAME就能用ResNet了？

HUI11126

我用RTX 2060 6G不太够用，即使跑一张40*40的图片也不行。我见2017年的FPN paper上面写的是使用8个GPU

HUI11126

Not found: ../data/tfrecord; No such file or directory

1

2020-12-29 11:04:21.313687: W tensorflow/core/framework/op_kernel.cc:1502] OP_REQUIRES failed at matching_files_op.cc:49 : Not found: ../data/tfrecord; No such file or directory 2020-12-29 11:04:21.313951: W tensorflow/core/framework/op_kernel.cc:1502] OP_REQUIRES failed at matching_files_op.cc:49 : Not found: ../data/tfrecord; No...

HUI11126

P6 = slim.avg_pool2d(P5, kernel_size=[1, 1], stride=2, scope='build_P6') 这个kernel_size不应该是[2, 2]么？

wanghangege


Iranb

他的权值全部自己初始化吗？不冻结网络全部训练

wangxi996

wangxi996

想不用预训练模型进行训练，应该怎样修改呢？ 

yts19871111

When I inference my data with the checkpoint obtained from the multi_gpu_train.py, it get an error as follow: Traceback (most recent call last): File "/root/anaconda3/envs/tensorflow12/lib/python3.6/site-packages/tensorflow/python/client/session.py", line 1334, in _do_call return...

XeChen

InvalidArgumentError (see above for traceback): Reshape cannot infer the missing input size for an empty tensor unless all specified input sizes are non-zero [[node Fast-RCNN/build_fc_layers/flatten_inputs/flatten/Reshape (defined at /home/dell-g3/anaconda3/lib/python3.6/site-packages/tensorflow/contrib/layers/python/layers/layers.py:1623) = Reshape[T=DT_FLOAT,...

dovesani