contextdesc icon indicating copy to clipboard operation
contextdesc copied to clipboard

Published 20 hours ago • verified publisher icon lzx551402

Please help: Get {'all_avg_recall': 0.0, 'v_avg_recall': 0.0, 'i_avg_recall': 0.0} from hseq_eval.py

Open GabbySuwichaya opened this issue 5 years ago • 2 comments

Thanks for releasing the code and I really enjoy reading your paper.

However, after I tried running hseq_eval.py with the following instruction:

python hseq_eval.py --function hseq_eval --config configs/hseq_eval.yaml

I received lots of warning which I am not sure if this is because of some incompatibility due to TF, or did I set something wrong.

I have installed all the prerequisite packages by with

$ cd contextdesc $ pip install -r requirements.txt

However, I could not find the packages for opencv-python==3.4.2.16 and opencv-contrib-python==3.4.2.16. So, I installed opencv-python==3.4.2.17 and opencv-contrib-python==3.4.2.17, instead.

So, just wonder if this is this ok? And is there anything else I have to install?

The full output messages look like this:

$ python hseq_eval.py --function hseq_eval --config configs/hseq_eval.yaml

WARNING:tensorflow:From /home/su-suwanwimolkul/contextdesc/utils/evaluator.py:26: The name tf.placeholder is deprecated. Please use tf.compat.v1.placeholder instead.

W0214 18:20:32.112720 140594310276864 deprecation_wrapper.py:119] From /home/su-suwanwimolkul/contextdesc/utils/evaluator.py:26: The name tf.placeholder is deprecated. Please use tf.compat.v1.placeholder instead.

`WARNING:tensorflow: The TensorFlow contrib module will not be included in TensorFlow 2.0. For more information, please see:

  • https://github.com/tensorflow/community/blob/master/rfcs/20180907-contrib-sunset.md
  • https://github.com/tensorflow/addons
  • https://github.com/tensorflow/io (for I/O related ops) If you depend on functionality not listed there, please file an issue.`

`W0214 18:20:32.400682 140594310276864 lazy_loader.py:50] The TensorFlow contrib module will not be included in TensorFlow 2.0. For more information, please see:

  • https://github.com/tensorflow/community/blob/master/rfcs/20180907-contrib-sunset.md
  • https://github.com/tensorflow/addons
  • https://github.com/tensorflow/io (for I/O related ops) If you depend on functionality not listed there, please file an issue.`

WARNING:tensorflow:From /home/su-suwanwimolkul/contextdesc/models/cnn_wrapper/network.py:225: conv2d (from tensorflow.python.layers.convolutional) is deprecated and will be removed in a future version. Instructions for updating: Use tf.keras.layers.Conv2Dinstead. W0214 18:20:32.400963 140594310276864 deprecation.py:323] From /home/su-suwanwimolkul/contextdesc/models/cnn_wrapper/network.py:225: conv2d (from tensorflow.python.layers.convolutional) is deprecated and will be removed in a future version. Instructions for updating: Usetf.keras.layers.Conv2Dinstead. WARNING:tensorflow:From /home/su-suwanwimolkul/anaconda3/envs/ContextDesc-Env/lib/python3.5/site-packages/tensorflow/python/ops/init_ops.py:1251: calling VarianceScaling.__init__ (from tensorflow.python.ops.init_ops) with dtype is deprecated and will be removed in a future version. Instructions for updating: Call initializer instance with the dtype argument instead of passing it to the constructor W0214 18:20:32.403049 140594310276864 deprecation.py:506] From /home/su-suwanwimolkul/anaconda3/envs/ContextDesc-Env/lib/python3.5/site-packages/tensorflow/python/ops/init_ops.py:1251: calling VarianceScaling.__init__ (from tensorflow.python.ops.init_ops) with dtype is deprecated and will be removed in a future version. Instructions for updating: Call initializer instance with the dtype argument instead of passing it to the constructor WARNING:tensorflow:From /home/su-suwanwimolkul/contextdesc/models/cnn_wrapper/network.py:403: batch_normalization (from tensorflow.python.layers.normalization) is deprecated and will be removed in a future version. Instructions for updating: Use keras.layers.BatchNormalization instead. In particular,tf.control_dependencies(tf.GraphKeys.UPDATE_OPS)should not be used (consult thetf.keras.layers.batch_normalizationdocumentation). W0214 18:20:32.574541 140594310276864 deprecation.py:323] From /home/su-suwanwimolkul/contextdesc/models/cnn_wrapper/network.py:403: batch_normalization (from tensorflow.python.layers.normalization) is deprecated and will be removed in a future version. Instructions for updating: Use keras.layers.BatchNormalization instead. In particular,tf.control_dependencies(tf.GraphKeys.UPDATE_OPS)should not be used (consult thetf.keras.layers.batch_normalizationdocumentation). WARNING:tensorflow:From /home/su-suwanwimolkul/contextdesc/models/cnn_wrapper/network.py:373: dense (from tensorflow.python.layers.core) is deprecated and will be removed in a future version. Instructions for updating: Use keras.layers.dense instead. W0214 18:20:32.843710 140594310276864 deprecation.py:323] From /home/su-suwanwimolkul/contextdesc/models/cnn_wrapper/network.py:373: dense (from tensorflow.python.layers.core) is deprecated and will be removed in a future version. Instructions for updating: Use keras.layers.dense instead. WARNING:tensorflow:From /home/su-suwanwimolkul/anaconda3/envs/ContextDesc-Env/lib/python3.5/site-packages/tensorflow/python/util/dispatch.py:180: batch_gather (from tensorflow.python.ops.array_ops) is deprecated and will be removed after 2017-10-25. Instructions for updating:tf.batch_gatheris deprecated, please usetf.gatherwithbatch_dimsinstead. W0214 18:20:33.143327 140594310276864 deprecation.py:323] From /home/su-suwanwimolkul/anaconda3/envs/ContextDesc-Env/lib/python3.5/site-packages/tensorflow/python/util/dispatch.py:180: batch_gather (from tensorflow.python.ops.array_ops) is deprecated and will be removed after 2017-10-25. Instructions for updating:tf.batch_gatheris deprecated, please usetf.gatherwithbatch_dims instead. 2020-02-14 18:20:33.349879: I tensorflow/core/platform/cpu_feature_guard.cc:142] Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX2 FMA 2020-02-14 18:20:33.373253: I tensorflow/core/platform/profile_utils/cpu_utils.cc:94] CPU Frequency: 4200000000 Hz 2020-02-14 18:20:33.373589: I tensorflow/compiler/xla/service/service.cc:168] XLA service 0x560ba56a4a70 executing computations on platform Host. Devices: 2020-02-14 18:20:33.373603: I tensorflow/compiler/xla/service/service.cc:175] StreamExecutor device (0): <undefined>, <undefined> WARNING:tensorflow:From /home/su-suwanwimolkul/anaconda3/envs/ContextDesc-Env/lib/python3.5/site-packages/tensorflow/python/training/saver.py:1276: checkpoint_exists (from tensorflow.python.training.checkpoint_management) is deprecated and will be removed in a future version. Instructions for updating: Use standard file APIs to check for files with this prefix. W0214 18:20:33.411734 140594310276864 deprecation.py:323] From /home/su-suwanwimolkul/anaconda3/envs/ContextDesc-Env/lib/python3.5/site-packages/tensorflow/python/training/saver.py:1276: checkpoint_exists (from tensorflow.python.training.checkpoint_management) is deprecated and will be removed in a future version. Instructions for updating: Use standard file APIs to check for files with this prefix. INFO:tensorflow:Restoring parameters from pretrained/contextdesc++/model.ckpt-400000 I0214 18:20:33.412194 140594310276864 saver.py:1280] Restoring parameters from pretrained/contextdesc++/model.ckpt-400000 2020-02-14 18:20:33.471508: W tensorflow/compiler/jit/mark_for_compilation_pass.cc:1412] (One-time warning): Not using XLA:CPU for cluster because envvar TF_XLA_FLAGS=--tf_xla_cpu_global_jit was not set. If you want XLA:CPU, either set that envvar, or use experimental_jit_scope to enable XLA:CPU. To confirm that XLA is active, pass --vmodule=xla_compilation_cache=1 (as a proper command-line flag, not via TF_XLA_FLAGS) or set the envvar XLA_FLAGS=--xla_hlo_profile.

``

GabbySuwichaya avatar Feb 14 '20 09:02 GabbySuwichaya

  1. Thank you for pointing out the wrong dependencies, and I have updated requirements accordingly.

  2. Those warning messages raised from tensorflow2.0 would not affect this program.

  3. Did you configure configs/hseq_eval.yaml correctly? To use the pre-extracted keypoints, please leave suffix to _cvpr. Otherwise you may simply leave it empty to evaluate from scratch. You will be able to see something like this:

I0215 12:15:02.382030 139998341805888 saver.py:1284] Restoring parameters from pretrained/contextdesc++/model.ckpt-400000
2020-02-15 12:15:06.992330: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcublas.so.10.0
2020-02-15 12:15:07.498549: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcudnn.so.7
0 v_abstract 0.8182284419682306
1 v_adam 0.8803200622186136
2 v_apprentices 0.9125884061453333
3 v_artisans 0.7532978162310019
4 v_astronautis 0.834236832593098
5 v_azzola 0.7433538034084249
6 v_bark 0.6238576094973912
7 v_bees 0.9546782656390992
8 v_beyus 0.6830293670376222
9 v_bip 0.8104020918066415
10 v_bird 0.7666802897475848
...

lzx551402 avatar Feb 15 '20 04:02 lzx551402

Thank you very much... Actually, the performance is very good.

Also, I would like to use your code on other datasets that also evaluate on different viewpoints and illumination. I will appropriately cite your work.

Since the code in image_matching.py provided in your repository is very easy to follow, I plan to use the image_matching.py to extract the features and save the extracted files.

Could you also provide the setting which could lead to the reported performance in HPatch dataset?

For example, how should I set up the following parameters in the image_matching.py :

  • n_sample
  • model_type
  • dense_desc
  • ratio_test
  • cross_check
  • reg_model_path
  • loc_model_path
  • aug_model_path

GabbySuwichaya avatar Feb 18 '20 08:02 GabbySuwichaya