1
$\begingroup$

I downloaded:

!git clone https://www.github.com/matterport/Mask_RCNN.git
os.chdir('Mask_RCNN')

And I've got an error: which version I should have of Keras?

WARNING:tensorflow:From /usr/local/lib/python3.5/dist-packages/keras/backend/tensorflow_backend.py:1154: calling reduce_max (from tensorflow.python.ops.math_ops) with keep_dims is deprecated and will be removed in a future version.
Instructions for updating:
keep_dims is deprecated, use keepdims instead
WARNING:tensorflow:From /usr/local/lib/python3.5/dist-packages/keras/backend/tensorflow_backend.py:1188: calling reduce_sum (from tensorflow.python.ops.math_ops) with keep_dims is deprecated and will be removed in a future version.
Instructions for updating:
keep_dims is deprecated, use keepdims instead
WARNING:tensorflow:From /usr/local/lib/python3.5/dist-packages/keras/backend/tensorflow_backend.py:1290: calling reduce_mean (from tensorflow.python.ops.math_ops) with keep_dims is deprecated and will be removed in a future version.
Instructions for updating:
keep_dims is deprecated, use keepdims instead

Futhermore:

totalMemory: 5.94GiB freeMemory: 5.44GiB
2019-04-03 22:37:38.374934: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1484] Adding visible gpu devices: 0, 1
2019-04-03 22:37:40.343417: I tensorflow/core/common_runtime/gpu/gpu_device.cc:965] Device interconnect StreamExecutor with strength 1 edge matrix:
2019-04-03 22:37:40.344366: I tensorflow/core/common_runtime/gpu/gpu_device.cc:971]      0 1 
2019-04-03 22:37:40.344373: I tensorflow/core/common_runtime/gpu/gpu_device.cc:984] 0:   N N 
2019-04-03 22:37:40.344377: I tensorflow/core/common_runtime/gpu/gpu_device.cc:984] 1:   N N 
2019-04-03 22:37:40.345556: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1097] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 11435 MB memory) -> physical GPU (device: 0, name: GeForce GTX TITAN X, pci bus id: 0000:02:00.0, compute capability: 5.2)
2019-04-03 22:37:40.450785: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1097] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:1 with 5220 MB memory) -> physical GPU (device: 1, name: GeForce GTX TITAN Black, pci bus id: 0000:01:00.0, compute capability: 3.5)
2019-04-03 22:37:42.518519: W tensorflow/core/framework/allocator.cc:108] Allocation of 51380224 exceeds 10% of system memory.
2019-04-03 22:37:42.601229: W tensorflow/core/framework/allocator.cc:108] Allocation of 51380224 exceeds 10% of system memory.
2019-04-03 22:37:51.648032: W tensorflow/core/framework/allocator.cc:108] Allocation of 51380224 exceeds 10% of system memory.
2019-04-03 22:37:51.678817: W tensorflow/core/framework/allocator.cc:108] Allocation of 51380224 exceeds 10% of system memory.
2019-04-03 22:37:51.706928: W tensorflow/core/framework/allocator.cc:108] Allocation of 51380224 exceeds 10% of system memory.
[I 22:37:55.611 NotebookApp] Starting buffering for fa2cd5ca-20f3-4472-b6ca-6821e2f56118:02508f46d629494ab46babe6d7611656
$\endgroup$

1 Answer 1

0
$\begingroup$

It's not an error, it's simply notifying you that the code is written using past versions of Tensorlfow and some of the arguments of special methods were going to be deprecated in the future releases of the library. It's ok to be used, but you can also check what version is used and make a virtual environment and install the specified version of the library you want.

You can run your cell twice in order not to see the warning again.

$\endgroup$
5
  • $\begingroup$ Ok. Futhermore my GPU is very slow. I used to use GTX 1050 and training one epoch last about 1-1.5h. $\endgroup$
    – Badum
    Apr 3, 2019 at 21:25
  • $\begingroup$ Now. I use two GPU (TITAN X, TITAN BLACK = 18GB) but time is similiar to GTX 1050(4GB). $\endgroup$
    – Badum
    Apr 3, 2019 at 21:26
  • $\begingroup$ You have to find your bottle neck. It can be somewhere between disk and memory or memory and gpu memory. $\endgroup$ Apr 3, 2019 at 21:33
  • $\begingroup$ How should I do it? $\endgroup$
    – Badum
    Apr 4, 2019 at 9:30
  • $\begingroup$ It is customary for DL tasks to load data to memory as much as possible and feed a subset of those to your gpu because gpu has a limited amount of memory. The first part can be implemented simply using generators. $\endgroup$ Apr 5, 2019 at 17:22

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.