Getting this trained so I could even start playing around was an absolute nightmare. Hopefully this will stop others from suffering the same pain.

Git is probably not the best way of distributing this as it's quite large (~30MB). Happy to follow up on any other suggestions.

just python main.py

a wrong result like below:

To construct input pipelines, use the `tf.data` module.
2019-04-14 14:37:44.570365: W tensorflow/core/framework/op_kernel.cc:1401] OP_REQUI2019-06-14 14:37:44.570725: W tensorflow/core/framework/op_kernel.cc:1401] OP_REQUIRES failed at whole_file_read_ops.cc:114 :
 Not found: NewRandomAccessFile failed to Create/Open: /SegNet/CamVid/trainannot/0016E5_05700.png : 系统找不到指定的路径。
; No such process
2019-04-14 14:37:44.631215: W tensorflow/core/framework/op_kernel.cc:1401] OP_REQUIRES failed at whole_file_read_ops.cc:114 : Not found: NewRand2019-06-14 14:37:44.631705: W tensorflow/core/framework/op_kerne
l.cc:1401] OP_REQUIRES failed at whole_file_read_ops.cc:114 : Not found: NewRandomAccessFile failed to Create/Open: /SegNet/CamVid/valannot/0016E5_08129.png : 系统找不到指定的路径。
; No such proces
```s

Hello,

Thank you for your hard work.

I am new to Tensorflow, so I am writing to ask for your help about check point file.

In the bash file, you mention about: "--finetune=/tmp3/first350/TensorFlow/Logs/model.ckpt-1000"

Could you please guide me how to get the model.ckpt?

Thank you so much for your help.

The code of drawings_object.writeImage is weird...

why not use:

def writeImage(image):
    """ store label data to colored image """
    plt.imshow(image, cmap=plt.get_cmap('flag'))
    

it is much simpler and without a fixed class types. It works fine to me.

I am trying to run the example.sh file and getting the following error. Can someone help?

The model is set to Training Max training Iteration: 20000 Initial lr: 0.001000 CamVid Image dir: /Users/username/Downloads/Tensorflow-SegNet-master/CamVid/train.txt CamVid Val dir: /Users/username/Downloads/Tensorflow-SegNet-master/CamVid/val.txt Batch Size: 5 Log dir: /Users/username/Downloads/Tensorflow-SegNet-master/Logs/--image_dir=/Users/username/Downloads/Tensorflow-SegNet-master/CamVid/train.txt Filling queue with 146 CamVid images before starting to train. This will take a few minutes. Filling queue with 146 CamVid images before starting to train. This will take a few minutes. Traceback (most recent call last): File "main.py", line 52, in tf.app.run() File "/Users/username/anaconda/lib/python3.6/site-packages/tensorflow/python/platform/app.py", line 48, in run _sys.exit(main(_sys.argv[:1] + flags_passthrough)) File "main.py", line 49, in main model.training(FLAGS, is_finetune=False) File "/Users/username/Downloads/Tensorflow-SegNet-master/model.py", line 369, in training loss, eval_prediction = inference(train_data_node, train_labels_node, batch_size, phase_train) File "/Users/username/Downloads/Tensorflow-SegNet-master/model.py", line 180, in inference conv1 = conv_layer_with_bn(norm1, [7, 7, images.get_shape().as_list()[3], 64], phase_train, name="conv1") File "/Users/username/Downloads/Tensorflow-SegNet-master/model.py", line 131, in conv_layer_with_bn conv_out = tf.nn.relu(batch_norm_layer(bias, train_phase, scope.name)) File "/Users/username/Downloads/Tensorflow-SegNet-master/model.py", line 171, in batch_norm_layer lambda: tf.contrib.layers.batch_norm(inputT, is_training=False, File "/Users/username/anaconda/lib/python3.6/site-packages/tensorflow/python/util/deprecation.py", line 296, in new_func return func(*args, **kwargs) File "/Users/username/anaconda/lib/python3.6/site-packages/tensorflow/python/ops/control_flow_ops.py", line 1819, in cond orig_res_t, res_t = context_t.BuildCondBranch(true_fn) File "/Users/username/anaconda/lib/python3.6/site-packages/tensorflow/python/ops/control_flow_ops.py", line 1694, in BuildCondBranch original_result = fn() File "/Users/username/Downloads/Tensorflow-SegNet-master/model.py", line 169, in lambda: tf.contrib.layers.batch_norm(inputT, is_training=True, File "/Users/username/anaconda/lib/python3.6/site-packages/tensorflow/python/util/lazy_loader.py", line 53, in getattr module = self._load() File "/Users/username/anaconda/lib/python3.6/site-packages/tensorflow/python/util/lazy_loader.py", line 42, in _load module = importlib.import_module(self.name) File "/Users/username/anaconda/lib/python3.6/importlib/init.py", line 126, in import_module return _bootstrap._gcd_import(name[level:], package, level) File "", line 978, in _gcd_import File "", line 961, in _find_and_load File "", line 950, in _find_and_load_unlocked File "", line 655, in _load_unlocked File "", line 678, in exec_module File "", line 205, in _call_with_frames_removed File "/Users/username/anaconda/lib/python3.6/site-packages/tensorflow/contrib/init.py", line 31, in from tensorflow.contrib import factorization File "/Users/username/anaconda/lib/python3.6/site-packages/tensorflow/contrib/factorization/init.py", line 24, in from tensorflow.contrib.factorization.python.ops.gmm import * File "/Users/username/anaconda/lib/python3.6/site-packages/tensorflow/contrib/factorization/python/ops/gmm.py", line 27, in from tensorflow.contrib.learn.python.learn.estimators import estimator File "/Users/username/anaconda/lib/python3.6/site-packages/tensorflow/contrib/learn/init.py", line 88, in from tensorflow.contrib.learn.python.learn import * File "/Users/username/anaconda/lib/python3.6/site-packages/tensorflow/contrib/learn/python/init.py", line 23, in from tensorflow.contrib.learn.python.learn import * File "/Users/username/anaconda/lib/python3.6/site-packages/tensorflow/contrib/learn/python/learn/init.py", line 25, in from tensorflow.contrib.learn.python.learn import estimators File "/Users/username/anaconda/lib/python3.6/site-packages/tensorflow/contrib/learn/python/learn/estimators/init.py", line 297, in from tensorflow.contrib.learn.python.learn.estimators.dnn import DNNClassifier File "/Users/username/anaconda/lib/python3.6/site-packages/tensorflow/contrib/learn/python/learn/estimators/dnn.py", line 30, in from tensorflow.contrib.learn.python.learn.estimators import dnn_linear_combined File "/Users/username/anaconda/lib/python3.6/site-packages/tensorflow/contrib/learn/python/learn/estimators/dnn_linear_combined.py", line 31, in from tensorflow.contrib.learn.python.learn.estimators import estimator File "/Users/username/anaconda/lib/python3.6/site-packages/tensorflow/contrib/learn/python/learn/estimators/estimator.py", line 49, in from tensorflow.contrib.learn.python.learn.learn_io import data_feeder File "/Users/username/anaconda/lib/python3.6/site-packages/tensorflow/contrib/learn/python/learn/learn_io/init.py", line 21, in from tensorflow.contrib.learn.python.learn.learn_io.dask_io import extract_dask_data File "/Users/username/anaconda/lib/python3.6/site-packages/tensorflow/contrib/learn/python/learn/learn_io/dask_io.py", line 26, in import dask.dataframe as dd File "/Users/username/anaconda/lib/python3.6/site-packages/dask/dataframe/init.py", line 5, in from .io import (from_array, from_pandas, from_bcolz, File "/Users/username/anaconda/lib/python3.6/site-packages/dask/dataframe/io/init.py", line 6, in from .csv import read_csv, to_csv, read_table File "/Users/username/anaconda/lib/python3.6/site-packages/dask/dataframe/io/csv.py", line 14, in from ...bytes import read_bytes File "/Users/username/anaconda/lib/python3.6/site-packages/dask/bytes/init.py", line 4, in from .core import read_bytes, open_files, open_text_files File "/Users/username/anaconda/lib/python3.6/site-packages/dask/bytes/core.py", line 9, in from .compression import seekable_files, files as compress_files File "/Users/username/anaconda/lib/python3.6/site-packages/dask/bytes/compression.py", line 32, in compress['snappy'] = snappy.compress AttributeError: module 'snappy' has no attribute 'compress'

Hi,

Thank you for sharing your code and helping us learn about segnet. 👍 . I had a small question.

In the model.py file (method: def cal_loss(logits, labels)), how is the array of the weights generated?

Thank you.

Hi, I just started exploring your codebase. I am trying to run segnet for my own project in which the images have fewer number of classes. When I try to just simply read one of the annotated images: '0001TP_006690.png' in the trainannot directory, I get this:

[[ 0.00392157 0.00392157 0.00392157 ..., 0.00392157 0.00392157 0.00392157] [ 0.00392157 0.00392157 0.00392157 ..., 0.00392157 0.00392157 0.00392157] [ 0.00392157 0.00392157 0.00392157 ..., 0.00392157 0.00392157 0.00392157] ..., [ 0.01568628 0.01568628 0.01568628 ..., 0.04313726 0.04313726 0.04313726] [ 0.01568628 0.01568628 0.01568628 ..., 0.04313726 0.04313726 0.04313726] [ 0.01568628 0.01568628 0.01568628 ..., 0.04313726 0.04313726 0.04313726]]

The way I tried to read the png is by: import matplotlib.image as img image = img.imread('0001TP_006690.png') print(image)

If i am not wrong, arent the classes supposed to be a whole number like 1,2..11(since there are 11 classes) instead of decimnal point numbers? or is there a conversion happening that I am not aware of.

Hi why the current code only stores checkpoint files of the latest 5000 steps? For example, if the upper limit of total steps is 20000, checkpoint files at steps of 1000, 2000, 3000, 4000,... first showed up and then disappeared as the training process proceeds. Finally we only get checkpoint files at steps of 16000, 17000, 18000, 19000, and 19999. Could we also store the checkpoint files at steps 1000, 2000, 3000, etc.?

2018-04-23 12:37:04.018255: step 110, loss = 1.55 (7.1 examples/sec; 0.281 sec/batch) accuracy = 0.999809 mean IU = 0.076908 class # 0 accuracy = 0.999809 class # 1 accuracy = 0.000000 class # 2 accuracy = 0.000000 class # 3 accuracy = 0.000000 class # 4 accuracy = 0.000000 class # 5 accuracy = 0.000000 class # 6 accuracy = 0.000000 class # 7 accuracy = 0.000000 class # 8 accuracy = 0.000000 class # 9 accuracy = 0.000000 class # 10 accuracy = 0.000000 class # 11 accuracy = 0.000000 class # 12 accuracy = 0.000000 class # 13 accuracy = 0.000000 class # 14 accuracy = 0.000000 class # 15 accuracy = 0.000000 class # 16 accuracy = 0.000000 class # 17 accuracy = 0.000000 class # 18 accuracy = 0.000000 class # 19 accuracy = 0.000000 class # 20 accuracy = 0.000000 class # 21 accuracy = 0.000000 class # 22 accuracy = 0.000000 class # 23 accuracy = 0.000000 class # 24 accuracy = 0.000000 class # 25 accuracy = 0.000000 class # 26 accuracy = 0.000000 [ 2 360 480 3] [ 2 360 480 1] [ 2 360 480 3] [ 2 360 480 1] [ 2 360 480 3] [ 2 360 480 1] [ 2 360 480 3] [ 2 360 480 1] [ 2 360 480 3] [ 2 360 480 1] [ 2 360 480 3] [ 2 360 480 1] [ 2 360 480 3] [ 2 360 480 1] [ 2 360 480 3] [ 2 360 480 1] [ 2 360 480 3] [ 2 360 480 1] [ 2 360 480 3] [ 2 360 480 1] 2018-04-23 12:37:07.865742: step 120, loss = 1.54 (7.1 examples/sec; 0.281 sec/batch) accuracy = 0.999881 mean IU = 0.076914 class # 0 accuracy = 0.999881 class # 1 accuracy = 0.000000 class # 2 accuracy = 0.000000 class # 3 accuracy = 0.000000 class # 4 accuracy = 0.000000 class # 5 accuracy = 0.000000 class # 6 accuracy = 0.000000 class # 7 accuracy = 0.000000 class # 8 accuracy = 0.000000 class # 9 accuracy = 0.000000 class # 10 accuracy = 0.000000 class # 11 accuracy = 0.000000 class # 12 accuracy = 0.000000 class # 13 accuracy = 0.000000 class # 14 accuracy = 0.000000 class # 15 accuracy = 0.000000 class # 16 accuracy = 0.000000 class # 17 accuracy = 0.000000 class # 18 accuracy = 0.000000 class # 19 accuracy = 0.000000 class # 20 accuracy = 0.000000 class # 21 accuracy = 0.000000 class # 22 accuracy = 0.000000 class # 23 accuracy = 0.000000 class # 24 accuracy = 0.000000 class # 25 accuracy = 0.000000 class # 26 accuracy = 0.000000 [ 2 360 480 3] [ 2 360 480 1] [ 2 360 480 3] [ 2 360 480 1] [ 2 360 480 3] [ 2 360 480 1] [ 2 360 480 3] [ 2 360 480 1] [ 2 360 480 3] [ 2 360 480 1] [ 2 360 480 3] [ 2 360 480 1] [ 2 360 480 3] [ 2 360 480 1] [ 2 360 480 3] [ 2 360 480 1] [ 2 360 480 3] [ 2 360 480 1] [ 2 360 480 3] [ 2 360 480 1] 2018-04-23 12:37:11.694385: step 130, loss = 1.53 (7.3 examples/sec; 0.273 sec/batch) accuracy = 1.000000 mean IU = 1.000000 class # 0 accuracy = 1.000000 class # 1 accuracy = 0.000000 class # 2 accuracy = 0.000000 class # 3 accuracy = 0.000000 class # 4 accuracy = 0.000000 class # 5 accuracy = 0.000000 class # 6 accuracy = 0.000000 class # 7 accuracy = 0.000000 class # 8 accuracy = 0.000000 class # 9 accuracy = 0.000000 class # 10 accuracy = 0.000000 class # 11 accuracy = 0.000000 class # 12 accuracy = 0.000000 class # 13 accuracy = 0.000000 class # 14 accuracy = 0.000000 class # 15 accuracy = 0.000000 class # 16 accuracy = 0.000000 class # 17 accuracy = 0.000000 class # 18 accuracy = 0.000000 class # 19 accuracy = 0.000000 class # 20 accuracy = 0.000000 class # 21 accuracy = 0.000000 class # 22 accuracy = 0.000000 class # 23 accuracy = 0.000000 class # 24 accuracy = 0.000000 class # 25 accuracy = 0.000000 class # 26 accuracy = 0.000000 [ 2 360 480 3] [ 2 360 480 1] [ 2 360 480 3] [ 2 360 480 1] [ 2 360 480 3] [ 2 360 480 1] [ 2 360 480 3] [ 2 360 480 1] [ 2 360 480 3] [ 2 360 480 1] [ 2 360 480 3] [ 2 360 480 1] [ 2 360 480 3] [ 2 360 480 1] [ 2 360 480 3] [ 2 360 480 1] [ 2 360 480 3] [ 2 360 480 1] [ 2 360 480 3] [ 2 360 480 1] 2018-04-23 12:37:15.534847: step 140, loss = 1.52 (7.4 examples/sec; 0.272 sec/batch) accuracy = 1.000000 mean IU = 1.000000 class # 0 accuracy = 1.000000 class # 1 accuracy = 0.000000 class # 2 accuracy = 0.000000 class # 3 accuracy = 0.000000 class # 4 accuracy = 0.000000 class # 5 accuracy = 0.000000 class # 6 accuracy = 0.000000 class # 7 accuracy = 0.000000 class # 8 accuracy = 0.000000 class # 9 accuracy = 0.000000 class # 10 accuracy = 0.000000 class # 11 accuracy = 0.000000 class # 12 accuracy = 0.000000 class # 13 accuracy = 0.000000 class # 14 accuracy = 0.000000 class # 15 accuracy = 0.000000 class # 16 accuracy = 0.000000 class # 17 accuracy = 0.000000 class # 18 accuracy = 0.000000 class # 19 accuracy = 0.000000 class # 20 accuracy = 0.000000 class # 21 accuracy = 0.000000 class # 22 accuracy = 0.000000 class # 23 accuracy = 0.000000 class # 24 accuracy = 0.000000 class # 25 accuracy = 0.000000 class # 26 accuracy = 0.000000 [ 2 360 480 3] [ 2 360 480 1] [ 2 360 480 3] [ 2 360 480 1] [ 2 360 480 3] [ 2 360 480 1] [ 2 360 480 3] [ 2 360 480 1] [ 2 360 480 3] [ 2 360 480 1] [ 2 360 480 3] [ 2 360 480 1] [ 2 360 480 3] [ 2 360 480 1] [ 2 360 480 3] [ 2 360 480 1] [ 2 360 480 3] [ 2 360 480 1] [ 2 360 480 3] [ 2 360 480 1] 2018-04-23 12:37:19.353283: step 150, loss = 1.51 (7.4 examples/sec; 0.272 sec/batch) accuracy = 0.999844

My project is to detect distress on road. Can this segnet tensorflow implementaion be used for that? If yes, then please guide me how to start? My dataset consists of 2k images with positive and negative samples. Would that be enough?