-
Hello, I am using the official resnet-101 pre-trained model (as your link). It is trained from ImageNet, with image order is RGB and IMAGE_MEAN is
```
_R_MEAN = 123.68 / 255
_G_MEAN = 116.78 / 255…
-
I trained the model (deeplab_resnet_init) with PASCAL VOC and only changed the number of classes to 2 (binary segmentation). Tested on my own dataset ang got mIoU = 0.69. How can I improve the results…
-
Thanks for the great project. I am using python 2.7 and tensorflow 1.2. I trained the code with `deeplab_resnet_init.ckpt` after 20k iterations:
`python train.py --random-mirror --random-scale`
Lo…
-
I used the pretrained Resnet-101 provided from the tensorflow as you told in 11/09/2017 update
(http://download.tensorflow.org/models/resnet_v1_101_2016_08_28.tar.gz), but an NotFoundError is raised …
-
Hi Nanqing,
First, thanks a lot for your implementation, this is a great piece of work!
I feel like you misunderstood the Multigrid block of DeepLabV3 network. You create a `bottleneck_hdc` unit…
-
Hello,
I have come upon an issue when running the train.py script, I assume it may be a bug due to me overlooking or miss-handling a step but for the life of me I can't figure out which :/
I run…
-
Thanks for sharing a nice work. I have achieved the Pascal VOC as you did. Right now, I would like to evaluate the performance in cityscape. I have created the training, validation and testing file as…
-
I am training your code with res101 in cityscape dataset (for deeplab pre-trained it worked well). I set up the `main.py` as
```
flags.DEFINE_float('momentum', 0.9, 'momentum')
flags.DEFINE_strin…
-
I tried your npy2ckpt.py to convert all the numpy weights I have and none of them can be converted successfully. They all come up with the message below:
> ValueError: Variable [_**depends on which …
-
As I try to train on the PASCAL VOC dataset, I received this error:
/……
s: 6.99GiB
I tensorflow/core/common_runtime/bfc_allocator.cc:702] Stats:
Limit: 7573333607
InUse: …