MarvinTeichmann/tensorflow-fcn
An Implementation of Fully Convolutional Networks in Tensorflow.
PythonMIT
Issues
- 0
- 4
Implementation for fully connected layer.
#27 opened by eveningdong - 7
can't download vgg16.npy file
#36 opened by jm-huang - 0
What the speed it can reach?
#59 opened by ming71 - 0
- 1
convert the ckpt model to movidius graph
#56 opened by zhanghanbin3159 - 2
A problem about fcn8_vgg.py
#43 opened by liuzhisheng1226 - 3
- 0
The weight decay (the regularization loss) is not applied for the fc_layer and upscore_layer
#58 opened by SpinachR - 0
Hello, could you tell me how to modify to achieve the effect in the article, you can run but the result is very poor.
#57 opened by linydf - 5
train just for segmentation
#29 opened by bhralzz - 1
- 0
Max Tensor Size in FCN-16 and FCN-32
#53 opened by aolsenjazz - 0
- 4
Training code
#26 opened by amirabs - 2
- 0
The problem about fcn8_vgg.py
#51 opened by SmmileIU - 0
- 3
A problem about the deconvolution
#46 opened by GZHermit - 0
- 0
A problem with fcn8_vgg.py
#47 opened by jionie - 0
- 0
Memory error
#41 opened by ztyxd - 0
- 3
AttributeError: type object 'GraphKeys' has no attribute 'REGULARIZATION_LOSSES'
#37 opened by FLSD-LEO - 2
Model initialization from scratch
#34 opened by aleksandrmelnyk - 1
ouput image doesn't make sense
#33 opened by aleksandrmelnyk - 1
- 0
why kSize of 64 in upscore_layer
#32 opened by yzhang123 - 0
Network seems not connected in Tensorboard
#30 opened by JayCHaos - 4
Error when run test_fcn32_vgg.py
#25 opened by Monsterhoho - 2
Problems with gray images
#24 opened by jiao0805 - 2
- 3
- 1
Change loss function to use tensorflow defaults
#22 opened by meetps - 4
running error from "python test_fcn32_vgg.py"
#20 opened by amiltonwong - 2
target_size=None can not work.
#16 opened by kuixu - 3
Error in loss function
#18 opened by ymzhang1919 - 1
Some errors in Fuinction of FCN32_vgg
#17 opened by rachit340 - 1
Why the memory of weights is 1.6G?
#15 opened by JackieLeeTHU11 - 1
Loss error
#13 opened by Yaffa1607 - 2
Possible bug in the upsampling layer?
#14 opened by patrickrbrandao - 6
- 4
training dataset
#10 opened by Yaffa1607 - 0
vgg.build() error : TypeError: build() missing 1 required positional argument: 'rgb'
#12 opened by Yaffa1607 - 1
General training code
#9 opened by wangyida - 1
fcn16: ValueError: Incompatible shapes for broadcasting: (?, ?, ?, 2) and (?, ?, ?, 20)
#8 opened by Thrasi - 3
training dataset
#6 opened by weiliu620 - 1
invalid syntax cross_entropy = -tf.reduce_sum(labels * tf.log(softmax), reduction_indices=[1]))
#4 opened by jiankang1991 - 0