huangh12

https://github.com/waspinator/pycococreator

huangh12

/*
 * Licensed to the Apache Software Foundation (ASF) under one
 * or more contributor license agreements.  See the NOTICE file
 * distributed with this work for additional information
 * regarding copyright ownership.  The ASF licenses this file
 * to you under the Apache License, Version 2.0 (the
 * "License"); you may not use this file except in compliance
 * with the License.  You may obtain a copy of the License at
 *
 *   http://www.apache.org/licenses/LICENSE-2.0

huangh12

filtered 10262 roidb entries: 28696 -> 18434
providing maximum shape [('data', (1, 3, 1024, 1024)), ('gt_boxes', (1, 100, 9))] [('label', (1, 36864)), ('bbox_target', (1, 36, 64, 646, 64, 64))]
{'bbox_target': (1L, 36L, 64L, 64L),
 'bbox_weight': (1L, 36L, 64L, 64L),
 'data': (1L, 3L, 1024L, 1024L),
 'gt_boxes': (1L, 5L, 9L),
 'im_info': (1L, 3L),
 'label': (1L, 36864L)}
('lr', 0.0005, 'lr_epoch_diff', [45.0, 52.0], 'lr_iters', [829530, 958568])
Epoch[0]Batch [50]Speed: 3.88 samples/sec   Train-RPNAcc=0.705270,   RPNLogLoss=0.590269,   RPNL1Loss=0.385927,   RCNNAcc=0.739890,   RCNL1Loss=0.106483,

huangh12

('classname:', 'plane')
('check fp:', array([ 0.,  0.,  0., ...,  1.,  1.,  1.]))
('check tp', array([ 1.,  1.,  1., ...,  0.,  0.,  0.]))
('npos num:', 2450)
('ap: ', 0.18839114907600038)
('classname:', 'baseball-diamond')
('check fp:', array([ 0.,  1.,  1.,  1.,  1.,  1.,  1.,  1.,  0.,  1.,  0.,  1.,  0.,
        1.,  0.,  1.,  0.,  1.,  1.,  1.,  0.,  1.,  0.,  0.,  1.,  1.,
        0.,  1.,  1.,  1.,  1.,  1.,  1.,  1.,  1.,  1.,  1.,  1.,  1.,
        1.,  0.,  1.,  0.,  0.,  0.,  1.,  0.,  1.,  1.,  1.,  1.,  1.,

huangh12

[----------] 2 tests from SigmoidCrossEntropyLossLayerTest/0, where TypeParam = caffe::CPUDevice<float>
[ RUN      ] SigmoidCrossEntropyLossLayerTest/0.TestSigmoidCrossEntropyLoss
src/caffe/test/test_sigmoid_cross_entropy_loss_layer.cpp:90: Failure
The difference between reference_loss and layer_loss is nan, which exceeds eps, where
reference_loss evaluates to 15.21638298034668,
layer_loss evaluates to -nan, and
eps evaluates to 0.019999999552965164.
debug: trial #0
src/caffe/test/test_sigmoid_cross_entropy_loss_layer.cpp:90: Failure
The difference between reference_loss and layer_loss is nan, which exceeds eps, where