./train.py error Synology DS1513+ DSM 6.1

matthewearl / deep-anpr

Using neural networks to build an automatic number plate recognition system

MIT License

1.84k stars 697 forks source link

./train.py error Synology DS1513+ DSM 6.1 #60

Closed CerBerE67112 closed 7 years ago

CerBerE67112 commented 7 years ago

Need help, thx.

Traceback (most recent call last): File "./train.py", line 266, in initial_weights=initial_weights) File "./train.py", line 174, in train digits_loss, presence_loss, loss = getloss(y, y) File "./train.py", line 132, in get_loss [-1, len(common.CHARS)])) File "/volume1/@appstore/py3k/usr/local/lib/python3.5/site-packages/tensorflow/python/ops/nn_ops.py", line 1555, in softmax_cross_entropy_with_logits labels, logits) File "/volume1/@appstore/py3k/usr/local/lib/python3.5/site-packages/tensorflow/python/ops/nn_ops.py", line 1510, in _ensure_xent_args "named arguments (labels=..., logits=..., ...)" % name) ValueError: Only call softmax_cross_entropy_with_logits with named arguments (labels=..., logits=..., ...)

longnguyenduy233 commented 7 years ago

This error due to the new version of tensorflow so you should change the line softmax_cross_entropy_with_logits to "softmax_cross_entropy_withlogits(labels= tf.reshape(y[:, 1:], [-1, len(common.CHARS)]),logits= tf.reshape(y[:, 1:], [-1, len(common.CHARS)]))" and the same for sigmoid.

CerBerE67112 commented 7 years ago

Thx @lovemoon23 I have done the correction on both... it'd seemed working... was in progress

Receiving just this alert...

WARNING:tensorflow:From /volume1/@appstore/py3k/usr/local/lib/python3.5/site-packages/tensorflow/python/util/tf_should_use.py:170: initialize_all_variables (from tensorflow.python.ops.variables) is deprecated and will be removed after 2017-03-02. Instructions for updating: Use tf.global_variables_initializer instead.

End then it returned

Killed

an idea why ?

longnguyenduy233 commented 7 years ago

It's just warning, i don't think it killed. Could you show me a stack trace??

CerBerE67112 commented 7 years ago

I'd love. Were can I find those traces ?

Maked the changes for the alert and received "Killed"...

# Depreciated in new version of tensorflow (https://github.com/tensorflow/tensorflow/blob/master/RELEASE.md#release-0120) 
# init = tf.initialize_all_variables()
# Replaced with
init = tf.global_variables_initializer()

CerBerE67112 commented 7 years ago

I also ran python3 -m trace --count -C . ./train.py

but did nothing else than returning a

Killed

after +/- 1 min.

CerBerE67112 commented 7 years ago

Thx @lovemoon23 you did the job. My server is not strong enough to run a training

dmesg | grep -E -i -B100 'killed'

[1737318.688401] [16606] 1024 16606 903682 250893 1407 121306 0 python3 [1737319.001678] Out of memory: Kill process 16606 (python3) score 278 or sacrifice child [1737325.126753] [16606] 1024 16606 903682 261508 1426 121708 0 python3 [1737325.432111] Out of memory: Kill process 16606 (python3) score 286 or sacrifice child [1737325.440058] Killed process 16606 (python3) total-vm:3614728kB, anon-rss:1046032kB, file-rss:0kB