Skip to content

Instantly share code, notes, and snippets.

https://github.com/waspinator/pycococreator
@huangh12
huangh12 / ordering_op-inl.h
Created March 18, 2020 17:50
add do_sort param for topk operator
/*
* 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
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,
('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
huangh12 / Mylog.LOG
Created June 16, 2017 13:05
make runtest for RCF
[----------] 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