deep-parking
deep-parking copied to clipboard
backend TkAgg version unknown
Hello Fabiocarrara, thank you for providing this project for everyone to try. At present, I have encountered a problem. I have an error after executing pklot.py. I also have this problem in main.py.
I also found debug information in the log, did not see any errors, only matplotlib will have a debug when just executed, how can this be solved?
matplotlib version 2.2.3
DEBUG 2019-02-11 00:28:43: CACHEDIR=/home/user/.cache/matplotlib
DEBUG 2019-02-11 00:28:43: Using fontManager instance from /home/user/.cache/matplotlib/fontList.json
DEBUG 2019-02-11 00:28:45: backend TkAgg version unknown
INFO 2019-02-11 00:28:45: Setting up mAlexNet trained on PKLot_UFPR05_train, validated on PKLot_UFPR04_train, PKLot_UFPR05_train, PKLot_PUC_train, tested on PKLot_UFPR04_test_PKLot_UFPR05_test_PKLot_PUC_test ...
INFO 2019-02-11 00:28:48: Training on PKLot_UFPR05_train while validating on PKLot_UFPR04_train, PKLot_UFPR05_train, PKLot_PUC_train ...
I0211 07:58:45.212375 2009 sgd_solver.cpp:112] Iteration 26166, lr = 0.0025
I0211 07:59:24.925103 2009 solver.cpp:239] Iteration 26215 (1.23388 iter/s, 39.712s/49 iters), loss = 0.00707085
I0211 07:59:24.925237 2009 solver.cpp:258] Train net output #0: loss = 0.00173975 (* 1 = 0.00173975 loss)
I0211 07:59:24.925256 2009 sgd_solver.cpp:112] Iteration 26215, lr = 0.0025
I0211 08:00:04.641950 2009 solver.cpp:239] Iteration 26264 (1.23376 iter/s, 39.716s/49 iters), loss = 0.00808991
I0211 08:00:04.642046 2009 solver.cpp:258] Train net output #0: loss = 0.000425289 (* 1 = 0.000425289 loss)
I0211 08:00:04.642061 2009 sgd_solver.cpp:112] Iteration 26264, lr = 0.0025
I0211 08:00:44.350703 2009 solver.cpp:239] Iteration 26313 (1.23401 iter/s, 39.708s/49 iters), loss = 0.00666064
I0211 08:00:44.350908 2009 solver.cpp:258] Train net output #0: loss = 0.00237676 (* 1 = 0.00237676 loss)
I0211 08:00:44.350924 2009 sgd_solver.cpp:112] Iteration 26313, lr = 0.0025
I0211 08:01:24.058185 2009 solver.cpp:239] Iteration 26362 (1.23404 iter/s, 39.707s/49 iters), loss = 0.008896
I0211 08:01:24.058348 2009 solver.cpp:258] Train net output #0: loss = 0.00142862 (* 1 = 0.00142862 loss)
I0211 08:01:24.058367 2009 sgd_solver.cpp:112] Iteration 26362, lr = 0.0025
I0211 08:02:03.743736 2009 solver.cpp:239] Iteration 26411 (1.23472 iter/s, 39.685s/49 iters), loss = 0.003344
I0211 08:02:03.743832 2009 solver.cpp:258] Train net output #0: loss = 0.000362561 (* 1 = 0.000362561 loss)
I0211 08:02:03.743847 2009 sgd_solver.cpp:112] Iteration 26411, lr = 0.0025
I0211 08:02:14.302935 2009 solver.cpp:464] Snapshotting to binary proto file snapshots/snapshot_iter_26425.caffemodel
I0211 08:02:14.303773 2009 sgd_solver.cpp:284] Snapshotting solver state to binary proto file snapshots/snapshot_iter_26425.solverstate
I0211 08:02:14.304250 2009 solver.cpp:332] Optimization Done.
I0211 08:02:14.304262 2009 caffe.cpp:250] Optimization Done.
DEBUG 2019-02-11 08:02:14: Summarizing mAlexNet-tr_PKLot_UFPR05_train-vl_PKLot_UFPR04_train_PKLot_UFPR05_train_PKLot_PUC_train-ts_PKLot_UFPR04_test_PKLot_UFPR05_test_PKLot_PUC_test ...
Traceback (most recent call last):
File "pklot.py", line 38, in
Thank you
Hi, for the first problem, you should check your matplotlib installation, try google: https://stackoverflow.com/questions/13110403/matplotlib-backend-missing-modules-with-underscore
About the IndexError, this arises when trying to read an empty experiment log. Check the logs of each experiment for errors during the training phase. It might be that one or more run have failed and contain no training log data. This makes the summarize() operation fail.