tf-image-segmentation

tf-image-segmentation copied to clipboard

fmthoker


huangzicheng

Set spatial_squeeze=False for resnet_v1 resnet_v1_101 and resnet_v1_50 models

LemonAniLabs

Incompatible shapes

12

I used import_meta_graph and restore and have used my own image. I am getting this error Incompatible shapes: [1,30,40,21] vs. [1,18,28,21] at image_np, pred_np = sess.run([image_tensor, pred], feed_dict=feed_dict_to_use)

vijtad

Getting widely different results on new runs.

1

Here are 3 results of segmentation on the same image.    Here is the relevant code. [test.py.zip](https://github.com/warmspringwinds/tf-image-segmentation/files/1081243/test.py.zip)

ankur2136

- supersedes and cleans up unnecessary files from #7. - Includes one command automated pascal_voc setup, plus examples of using keras tensors. - Also includes integration of keras layers with...

ahundt

dataset for convert_pascal_voc_to_tfrecords.ipynb

2

Hi, @warmspringwinds , According to this [notebook](https://github.com/warmspringwinds/tf-image-segmentation/blob/master/tf_image_segmentation/recipes/pascal_voc/convert_pascal_voc_to_tfrecords.ipynb), the dataset for VOC2012 and benchmark_RELEASE are needed? Could you list the download link for the corresponding dataset? Thanks!

amiltonwong

can not find "fc_conv_padding" parameter in function vgg.vgg_16

1

when i read the code fcn_32s.py , i find that it called function vgg.vgg_16 and passing into a parameter called fc_conv_padding="SAME". logits, end_points = vgg.vgg_16(mean_centered_image_batch, num_classes=number_of_classes, is_training=is_training, spatial_squeeze=False, fc_conv_padding='SAME') However,...

ghost

Fix for #8: make forward/backward compatible with tf-1.0.x.

8

Fix for #8 based on comments by @vijtad and @zhaozj89. Add version checks/changes at fail points based on output of tf_upgrade_tool.py. Add Python 3.x compatibility. Unsure if adding forward/backward compatibility...

barryridge

Does it work in tensorflow v1.0.0?

2

It is a great library. But when I use tensorflow v1.0.0 it didn't work. My code: class FcvImgSeg(object): def __init__(self): print('init') self.slim = tf.contrib.slim self.fcv_ckpt_path = './models/lcct/model_fcn8s_final.ckpt' self.number_of_classes = 21...

yt7589