luminoth issues

Bump tensorflow from 1.5.0 to 2.7.2 in /docs

Bumps [tensorflow](https://github.com/tensorflow/tensorflow) from 1.5.0 to 2.7.2. Release notes Sourced from tensorflow's releases. TensorFlow 2.7.2 Release 2.7.2 This releases introduces several vulnerability fixes: Fixes a code injection in saved_model_cli (CVE-2022-29216) Fixes...

dependabot[bot]

dependencies

Performance issue in /luminoth/eval.py (by P3)

1

Hello! I've found a performance issue in /luminoth/eval.py: `with tf.Session() as sess`[(here)](https://github.com/tryolabs/luminoth/blob/9109d8b98bcbb0ad4e8e0c928da0b20627fa774f/luminoth/eval.py#L312) is defined in the function `evaluate_once`[(here)](https://github.com/tryolabs/luminoth/blob/9109d8b98bcbb0ad4e8e0c928da0b20627fa774f/luminoth/eval.py#L282) which is repeatedly called in the loop `for checkpoint in checkpoints`[(here)](https://github.com/tryolabs/luminoth/blob/9109d8b98bcbb0ad4e8e0c928da0b20627fa774f/luminoth/eval.py#L193). `tf.Session`...

DLPerf

Height and Width Reversed in crop_and_resize

On the following lines, a list containing two times the pooled width and two times the pooled height (in that order) is provided to tf.image.crop_and_resize: https://github.com/tryolabs/luminoth/blob/9109d8b98bcbb0ad4e8e0c928da0b20627fa774f/luminoth/models/fasterrcnn/roi_pool.py#L75-L78 However, in the [crop_and_resize...

leland-hepworth

docs: fix simple typo, targers -> targets

There is a small typo in luminoth/models/fasterrcnn/rpn_target.py. Should read `targets` rather than `targers`. Semi-automated pull request generated by https://github.com/timgates42/meticulous/blob/master/docs/NOTE.md

timgates42

INFO:tensorflow:Error reported to Coordinator: <class 'tensorflow.python.framework.errors_impl.InvalidArgumentError'>, Name: , Context feature 'depth' is required but could not be found.

3

I'm getting above error while training fasterrcnn using luminoth. can someone provide solution for this. Error:- tensorflow.python.framework.errors_impl.InvalidArgumentError: Name: , Context feature 'depth' is required but could not be found. [[Node:...

ghost

Type: Needs Investigation

Type: Bug

Traceback (most recent call last): File "d:\anaconda3-2019.03\anaconda3\envs\luminoth\lib\site-packages\urllib3\connection.py", line 160, in _new_conn (self._dns_host, self.port), self.timeout, **extra_kw File "d:\anaconda3-2019.03\anaconda3\envs\luminoth\lib\site-packages\urllib3\util\connection.py", line 61, in create_connection for res in socket.getaddrinfo(host, port, family, socket.SOCK_STREAM): File "d:\anaconda3-2019.03\anaconda3\envs\luminoth\lib\socket.py", line 743, in getaddrinfo for res in _socket.getaddrinfo(host, port, family, type, proto, flags): socket.gaierror: [Errno 11001] getaddrinfo failed

When Training Faster RCNN, I got this problem! Traceback (most recent call last): File "d:\anaconda3-2019.03\anaconda3\envs\luminoth\lib\site-packages\urllib3\connection.py", line 160, in _new_conn (self._dns_host, self.port), self.timeout, **extra_kw File "d:\anaconda3-2019.03\anaconda3\envs\luminoth\lib\site-packages\urllib3\util\connection.py", line 61, in create_connection for...

BlackHandguy

Add a safe wrapper for cross entropy.

This fixes an error we were getting when `tf.nn.softmax_cross_entropy_with_logits` received empty tensors. The error was quite elusive, as it only happened on GPUs and not CPUs. [This](https://github.com/tensorflow/tensorflow/issues/6766) is the issue...

IanTayler