search

toni-heittola / dcase2019_task1_baseline

DCASE2019 Challenge Task 1 baseline system
MIT License
20 stars 11 forks source link

Task 1c fails to produce leaderboard results #3

Open bernhardlehner opened 5 years ago

bernhardlehner commented 5 years ago

I'm not sure if this is officially supported already, but since there are leaderboard results at the Kaggle site, I suppose it should work. Default mode seems to work as expected. I used the latest version and setup a new environment from scratch, just to be on the safe side.

The download of the leaderboard dataset did not start, and I tried to add it manually to the default path, but to no avail. I attached the log, maybe this helps.

OS: WIN10 Environment: _tflow_select 2.1.0 gpu absl-py 0.7.1 py37_0 astor 0.7.1 py37_0 attrs 19.1.0 py37_1 audioread 2.1.8 pypi_0 pypi backcall 0.1.0 py37_0 blas 1.0 mkl bleach 3.1.0 py37_0 ca-certificates 2019.1.23 0 certifi 2019.3.9 py37_0 cffi 1.12.3 pypi_0 pypi chardet 3.0.4 pypi_0 pypi colorama 0.4.1 py37_0 cudatoolkit 10.0.130 0 cudnn 7.3.1 cuda10.0_0 cycler 0.10.0 pypi_0 pypi dcase-util 0.2.10 pypi_0 pypi decorator 4.4.0 py37_1 defusedxml 0.6.0 py_0 entrypoints 0.3 py37_0 future 0.17.1 pypi_0 pypi gast 0.2.2 py37_0 grpcio 1.16.1 py37h351948d_1 h5py 2.9.0 py37h5e291fa_0 hdf5 1.10.4 h7ebc959_0 icc_rt 2019.0.0 h0cc432a_1 icu 58.2 ha66f8fd_1 idna 2.8 pypi_0 pypi intel-openmp 2019.3 203 ipykernel 5.1.0 py37h39e3cac_0 ipython 7.5.0 py37h39e3cac_0 ipython_genutils 0.2.0 py37_0 ipywidgets 7.4.2 py37_0 jedi 0.13.3 py37_0 jinja2 2.10.1 py37_0 joblib 0.13.2 pypi_0 pypi jpeg 9b hb83a4c4_2 jsonschema 3.0.1 py37_0 jupyter 1.0.0 py37_7 jupyter_client 5.2.4 py37_0 jupyter_console 6.0.0 py37_0 jupyter_core 4.4.0 py37_0 keras 2.2.4 pypi_0 pypi keras-applications 1.0.7 py_0 keras-preprocessing 1.0.9 py_0 kiwisolver 1.1.0 pypi_0 pypi libpng 1.6.37 h2a8f88b_0 libprotobuf 3.7.1 h7bd577a_0 librosa 0.6.3 pypi_0 pypi libsodium 1.0.16 h9d3ae62_0 llvmlite 0.29.0 pypi_0 pypi m2w64-gcc-libgfortran 5.3.0 6 m2w64-gcc-libs 5.3.0 7 m2w64-gcc-libs-core 5.3.0 7 m2w64-gmp 6.1.0 2 m2w64-libwinpthread-git 5.0.0.4634.697f757 2 markdown 3.1 py37_0 markupsafe 1.1.1 py37he774522_0 matplotlib 3.1.0 pypi_0 pypi mistune 0.8.4 py37he774522_0 mkl 2019.3 203 mkl_fft 1.0.12 py37h14836fe_0 mkl_random 1.0.2 py37h343c172_0 mock 3.0.5 py37_0 msys2-conda-epoch 20160418 1 nbconvert 5.5.0 py_0 nbformat 4.4.0 py37_0 notebook 5.7.8 py37_0 numba 0.44.0 pypi_0 pypi numpy 1.16.4 py37h19fb1c0_0 numpy-base 1.16.4 py37hc3f5095_0 openssl 1.1.1c he774522_1 pandoc 2.2.3.2 0 pandocfilters 1.4.2 py37_1 parso 0.4.0 py_0 pickleshare 0.7.5 py37_0 pip 19.1.1 py37_0 prometheus_client 0.6.0 py37_0 prompt_toolkit 2.0.9 py37_0 protobuf 3.7.1 py37h33f27b4_0 pycparser 2.19 pypi_0 pypi pydot-ng 2.0.0 pypi_0 pypi pygments 2.4.0 py_0 pyparsing 2.4.0 pypi_0 pypi pyqt 5.9.2 py37h6538335_2 pyreadline 2.1 py37_1 pyrsistent 0.14.11 py37he774522_0 python 3.7.3 h8c8aaf0_1 python-dateutil 2.8.0 py37_0 python-magic 0.4.15 pypi_0 pypi pywinpty 0.5.5 py37_1000 pyyaml 5.1 pypi_0 pypi pyzmq 18.0.0 py37ha925a31_0 qt 5.9.7 vc14h73c81de_0 qtconsole 4.5.1 py_0 requests 2.22.0 pypi_0 pypi resampy 0.2.1 pypi_0 pypi scikit-learn 0.21.2 pypi_0 pypi scipy 1.2.1 py37h29ff71c_0 sed-eval 0.2.1 pypi_0 pypi send2trash 1.5.0 py37_0 setuptools 41.0.1 py37_0 sip 4.19.8 py37h6538335_0 six 1.12.0 py37_0 soundfile 0.10.2 pypi_0 pypi sqlite 3.28.0 he774522_0 tensorboard 1.13.1 py37h33f27b4_0 tensorflow 1.13.1 gpu_py37h83e5d6a_0 tensorflow-base 1.13.1 gpu_py37h871c8ca_0 tensorflow-estimator 1.13.0 py_0 tensorflow-gpu 1.13.1 h0d30ee6_0 termcolor 1.1.0 py37_1 terminado 0.8.2 py37_0 testpath 0.4.2 py37_0 tornado 6.0.2 py37he774522_0 tqdm 4.32.1 pypi_0 pypi traitlets 4.3.2 py37_0 urllib3 1.25.3 pypi_0 pypi validators 0.13.0 pypi_0 pypi vc 14.1 h0510ff6_4 vs2015_runtime 14.15.26706 h3a45250_4 wcwidth 0.1.7 py37_0 webencodings 0.5.1 py37_1 werkzeug 0.15.2 py_0 wheel 0.33.4 py37_0 widgetsnbextension 3.4.2 py37_0 wincertstore 0.2 py37_0 winpty 0.4.3 4 zeromq 4.3.1 h33f27b4_3 zlib 1.2.11 h62dcd97_3

task1a.log

toni-heittola commented 5 years ago

Based on attached is task1a log there seems to be some error happening inside tensorflow. Based on package versions, you could try to solve this by downgrading tensorflow (1.9.0) and keras (2.2.2) versions slightly. System has been developed with versions defined in requirements.txt and in Linux.

Leaderboard and Evaluation datasets should be downloaded automatically in 'leaderboard' or 'eval' modes, however, before that system will train a model with all development dataset, and it seems that based on the log this stage is giving you an error.

bernhardlehner commented 5 years ago

Thank you for your suggestion, I appreciate it. However, I am still not able to get the leaderboard output (output file is created, but it only contains the header). After creating a new environment with the suggested versions of TF and KERAS from scratch, the leaderboard data is downloaded, the features are extracted, but then it seems that some GPU property cannot be obtained...here is the cmd line output (I tried it twice with --overwrite, ans this is reproducable): ... [I] [ 1196 / 1200] [996.wav] [I] [ 1197 / 1200] [997.wav] [I] [ 1198 / 1200] [998.wav] [I] [ 1199 / 1200] [999.wav] [I] DONE [0:00:48.691899] [1200 items] [0:00:00.040577 per item] [I] [I] Testing [I] ======================================== [I] Fold [all_data] 2019-06-01 19:30:14.945024: E C:\users\nwani_bazel_nwani\ujdkfsks\execroot\org_tensorflow\tensorflow\core\grappler\clusters\utils.cc:82] Failed to get device properties, error code: 30 [I] DONE [0:00:02.070353] [I] [I] System output saved to [output.csv]

What puzzles me is the path that you can see in the above output, because this is not my home directory...could it be that it has something to do with that?