A Classification Problem with Trained Model

[LawBow]

I build a 250 categories training dataset, each category has more than 10k images. Following the Imagenet training steps, I got 92% accuracy in test.

I1101 21:19:04.669459 19693 solver.cpp:247] Iteration 13000, Testing net (#0)
I1101 21:20:40.372087 19693 solver.cpp:298]     Test net output #0: accuracy = 0.92135
I1101 21:20:40.372143 19693 solver.cpp:298]     Test net output #1: loss = 0.290574 (* 1 = 0.290574 loss)
I1101 21:20:41.380187 19693 solver.cpp:191] Iteration 13000, loss = 0.310368
I1101 21:20:41.380216 19693 solver.cpp:206]     Train net output #0: loss = 0.310368 (* 1 = 0.310368 loss)
I1101 21:20:41.380236 19693 solver.cpp:403] Iteration 13000, lr = 0.01

I use such trained model and test valid dataset. the valid dataset random selected from database, so the data distribution is the same with training and test data. I also subtract the trained dataset mean file. But the test result is 44%, I don't know what's the problem?

net = caffe.Classifier(MODEL_FILE, PRETRAINED,
                       mean=np.load(mean_file),
                       raw_scale=255,
                       image_dims=(64, 64))

[dabilied]

overfit,simpler net may work

[dabilied]

oh sorry ,if you use valid dataset for test in training , the problem seems just occurs during testing procedure,you can try to test in C++.

[PatWie]

You validation set cannot have the same distributions. Errors can be:

• double subtracting the mean image
• wrong image dimensions (wrong part of the image for testing)
• wrong model (double check path)?
• input_dim in your deploy-net-layout

I do not use the python interface for some reasons. You can try to print the prediction within caffe. Adjust the solver.prototxt:

• set learning_rate=0
• set max_iter=1
• change path to your validation leveldb

And instead of the softmax__loss_ layer in your train_val.prototxt:

layers {
  name: "loss"
  type: SOFTMAX_LOSS
  bottom: "full_layer_002"
  bottom: "label"
  top: "loss"
}

you can use a softmax-layer and:
https://github.com/PatWie/caffe/blob/teststatistic/src/caffe/layers/statistic_layer.cpp
which simply prints out the predictions.

layers {
  name: "prob"
  type: SOFTMAX
  bottom: "full_layer_002"
  top: "prob"
}
layers {
  name: "stats"
  type: TESTSTATISTIC
  bottom: "prob"
  bottom: "label"
  top: "stats"
}

The output will be

// layout of line
[>>] true_label prob_0 label_0 prob_1 label_1 prob_2 label_2 prob_3 label_3

use

caffe train --solver=YOUR_SOLVER --weights=YOUR_MODEL &> run_plot.log

Then simply

grep "\[>>\]" run_plot.log | tail -n 800 |  sed 's/\[>>\] //g'  > prediction_test.csv

if your validation set has 800 images. You may compute the correct error from there.

[tjusxh]

@LawBow Hi,
Could you help me?I come across one question which there are 160 dimension in the full connection layer and after softmax 2000+ categories.Could it work?
Thanks
tjusxh

[shelhamer]

@PatWie's summary is good. These problems will go away once #1245 is addressed.

[sulthanashafi]

iam also in the similar situation i tried changing augmentation as mentioned but haven't worked .Please someone help solve it.i may attach the few lkstlines of log file here.I0413 06:08:56.441606 22206 net.cpp:159] Memory required for data: 5099532756
I0413 06:08:56.441609 22206 net.cpp:222] mAP does not need backward computation.
I0413 06:08:56.441613 22206 net.cpp:222] score does not need backward computation.
I0413 06:08:56.441617 22206 net.cpp:222] cluster_gt does not need backward computation.
I0413 06:08:56.441622 22206 net.cpp:222] cluster does not need backward computation.
I0413 06:08:56.441649 22206 net.cpp:220] coverage_loss needs backward computation.
I0413 06:08:56.441654 22206 net.cpp:220] bbox_loss needs backward computation.
I0413 06:08:56.441659 22206 net.cpp:220] bbox-obj-norm needs backward computation.
I0413 06:08:56.441663 22206 net.cpp:220] bbox-norm needs backward computation.
I0413 06:08:56.441668 22206 net.cpp:220] bbox_mask needs backward computation.
I0413 06:08:56.441673 22206 net.cpp:220] bboxes_bbox/regressor_0_split needs backward computation.
I0413 06:08:56.441678 22206 net.cpp:220] bbox/regressor needs backward computation.
I0413 06:08:56.441682 22206 net.cpp:220] coverage_coverage/sig_0_split needs backward computation.
I0413 06:08:56.441686 22206 net.cpp:220] coverage/sig needs backward computation.
I0413 06:08:56.441690 22206 net.cpp:220] cvg/classifier needs backward computation.
I0413 06:08:56.441695 22206 net.cpp:220] pool5/drop_s1_pool5/drop_s1_0_split needs backward computation.
I0413 06:08:56.441699 22206 net.cpp:220] pool5/drop_s1 needs backward computation.
I0413 06:08:56.441704 22206 net.cpp:220] inception_5b/output needs backward computation.
I0413 06:08:56.441709 22206 net.cpp:220] inception_5b/relu_pool_proj needs backward computation.
I0413 06:08:56.441712 22206 net.cpp:220] inception_5b/pool_proj needs backward computation.
I0413 06:08:56.441716 22206 net.cpp:220] inception_5b/pool needs backward computation.
I0413 06:08:56.441721 22206 net.cpp:220] inception_5b/relu_5x5 needs backward computation.
I0413 06:08:56.441725 22206 net.cpp:220] inception_5b/5x5 needs backward computation.
I0413 06:08:56.441730 22206 net.cpp:220] inception_5b/relu_5x5_reduce needs backward computation.
I0413 06:08:56.441733 22206 net.cpp:220] inception_5b/5x5_reduce needs backward computation.
I0413 06:08:56.441737 22206 net.cpp:220] inception_5b/relu_3x3 needs backward computation.
I0413 06:08:56.441741 22206 net.cpp:220] inception_5b/3x3 needs backward computation.
I0413 06:08:56.441745 22206 net.cpp:220] inception_5b/relu_3x3_reduce needs backward computation.
I0413 06:08:56.441750 22206 net.cpp:220] inception_5b/3x3_reduce needs backward computation.
I0413 06:08:56.441753 22206 net.cpp:220] inception_5b/relu_1x1 needs backward computation.
I0413 06:08:56.441756 22206 net.cpp:220] inception_5b/1x1 needs backward computation.
I0413 06:08:56.441761 22206 net.cpp:220] inception_5a/output_inception_5a/output_0_split needs backward computation.
I0413 06:08:56.441766 22206 net.cpp:220] inception_5a/output needs backward computation.
I0413 06:08:56.441771 22206 net.cpp:220] inception_5a/relu_pool_proj needs backward computation.
I0413 06:08:56.441774 22206 net.cpp:220] inception_5a/pool_proj needs backward computation.
I0413 06:08:56.441778 22206 net.cpp:220] inception_5a/pool needs backward computation.
I0413 06:08:56.441782 22206 net.cpp:220] inception_5a/relu_5x5 needs backward computation.
I0413 06:08:56.441787 22206 net.cpp:220] inception_5a/5x5 needs backward computation.
I0413 06:08:56.441790 22206 net.cpp:220] inception_5a/relu_5x5_reduce needs backward computation.
I0413 06:08:56.441794 22206 net.cpp:220] inception_5a/5x5_reduce needs backward computation.
I0413 06:08:56.441798 22206 net.cpp:220] inception_5a/relu_3x3 needs backward computation.
I0413 06:08:56.441802 22206 net.cpp:220] inception_5a/3x3 needs backward computation.
I0413 06:08:56.441805 22206 net.cpp:220] inception_5a/relu_3x3_reduce needs backward computation.
I0413 06:08:56.441809 22206 net.cpp:220] inception_5a/3x3_reduce needs backward computation.
I0413 06:08:56.441813 22206 net.cpp:220] inception_5a/relu_1x1 needs backward computation.
I0413 06:08:56.441817 22206 net.cpp:220] inception_5a/1x1 needs backward computation.
I0413 06:08:56.441820 22206 net.cpp:220] inception_4e/output_inception_4e/output_0_split needs backward computation.
I0413 06:08:56.441825 22206 net.cpp:220] inception_4e/output needs backward computation.
I0413 06:08:56.441830 22206 net.cpp:220] inception_4e/relu_pool_proj needs backward computation.
I0413 06:08:56.441834 22206 net.cpp:220] inception_4e/pool_proj needs backward computation.
I0413 06:08:56.441844 22206 net.cpp:220] inception_4e/pool needs backward computation.
I0413 06:08:56.441848 22206 net.cpp:220] inception_4e/relu_5x5 needs backward computation.
I0413 06:08:56.441853 22206 net.cpp:220] inception_4e/5x5 needs backward computation.
I0413 06:08:56.441856 22206 net.cpp:220] inception_4e/relu_5x5_reduce needs backward computation.
I0413 06:08:56.441860 22206 net.cpp:220] inception_4e/5x5_reduce needs backward computation.
I0413 06:08:56.441864 22206 net.cpp:220] inception_4e/relu_3x3 needs backward computation.
I0413 06:08:56.441867 22206 net.cpp:220] inception_4e/3x3 needs backward computation.
I0413 06:08:56.441871 22206 net.cpp:220] inception_4e/relu_3x3_reduce needs backward computation.
I0413 06:08:56.441875 22206 net.cpp:220] inception_4e/3x3_reduce needs backward computation.
I0413 06:08:56.441879 22206 net.cpp:220] inception_4e/relu_1x1 needs backward computation.
I0413 06:08:56.441884 22206 net.cpp:220] inception_4e/1x1 needs backward computation.
I0413 06:08:56.441887 22206 net.cpp:220] inception_4d/output_inception_4d/output_0_split needs backward computation.
I0413 06:08:56.441891 22206 net.cpp:220] inception_4d/output needs backward computation.
I0413 06:08:56.441896 22206 net.cpp:220] inception_4d/relu_pool_proj needs backward computation.
I0413 06:08:56.441900 22206 net.cpp:220] inception_4d/pool_proj needs backward computation.
I0413 06:08:56.441905 22206 net.cpp:220] inception_4d/pool needs backward computation.
I0413 06:08:56.441908 22206 net.cpp:220] inception_4d/relu_5x5 needs backward computation.
I0413 06:08:56.441912 22206 net.cpp:220] inception_4d/5x5 needs backward computation.
I0413 06:08:56.441916 22206 net.cpp:220] inception_4d/relu_5x5_reduce needs backward computation.
I0413 06:08:56.441920 22206 net.cpp:220] inception_4d/5x5_reduce needs backward computation.
I0413 06:08:56.441925 22206 net.cpp:220] inception_4d/relu_3x3 needs backward computation.
I0413 06:08:56.441927 22206 net.cpp:220] inception_4d/3x3 needs backward computation.
I0413 06:08:56.441931 22206 net.cpp:220] inception_4d/relu_3x3_reduce needs backward computation.
I0413 06:08:56.441934 22206 net.cpp:220] inception_4d/3x3_reduce needs backward computation.
I0413 06:08:56.441938 22206 net.cpp:220] inception_4d/relu_1x1 needs backward computation.
I0413 06:08:56.441942 22206 net.cpp:220] inception_4d/1x1 needs backward computation.
I0413 06:08:56.441946 22206 net.cpp:220] inception_4c/output_inception_4c/output_0_split needs backward computation.
I0413 06:08:56.441951 22206 net.cpp:220] inception_4c/output needs backward computation.
I0413 06:08:56.441956 22206 net.cpp:220] inception_4c/relu_pool_proj needs backward computation.
I0413 06:08:56.441958 22206 net.cpp:220] inception_4c/pool_proj needs backward computation.
I0413 06:08:56.441962 22206 net.cpp:220] inception_4c/pool needs backward computation.
I0413 06:08:56.441967 22206 net.cpp:220] inception_4c/relu_5x5 needs backward computation.
I0413 06:08:56.441969 22206 net.cpp:220] inception_4c/5x5 needs backward computation.
I0413 06:08:56.441973 22206 net.cpp:220] inception_4c/relu_5x5_reduce needs backward computation.
I0413 06:08:56.441977 22206 net.cpp:220] inception_4c/5x5_reduce needs backward computation.
I0413 06:08:56.441982 22206 net.cpp:220] inception_4c/relu_3x3 needs backward computation.
I0413 06:08:56.441985 22206 net.cpp:220] inception_4c/3x3 needs backward computation.
I0413 06:08:56.441988 22206 net.cpp:220] inception_4c/relu_3x3_reduce needs backward computation.
I0413 06:08:56.441992 22206 net.cpp:220] inception_4c/3x3_reduce needs backward computation.
I0413 06:08:56.441997 22206 net.cpp:220] inception_4c/relu_1x1 needs backward computation.
I0413 06:08:56.441999 22206 net.cpp:220] inception_4c/1x1 needs backward computation.
I0413 06:08:56.442003 22206 net.cpp:220] inception_4b/output_inception_4b/output_0_split needs backward computation.
I0413 06:08:56.442008 22206 net.cpp:220] inception_4b/output needs backward computation.
I0413 06:08:56.442013 22206 net.cpp:220] inception_4b/relu_pool_proj needs backward computation.
I0413 06:08:56.442020 22206 net.cpp:220] inception_4b/pool_proj needs backward computation.
I0413 06:08:56.442024 22206 net.cpp:220] inception_4b/pool needs backward computation.
I0413 06:08:56.442028 22206 net.cpp:220] inception_4b/relu_5x5 needs backward computation.
I0413 06:08:56.442032 22206 net.cpp:220] inception_4b/5x5 needs backward computation.
I0413 06:08:56.442036 22206 net.cpp:220] inception_4b/relu_5x5_reduce needs backward computation.
I0413 06:08:56.442039 22206 net.cpp:220] inception_4b/5x5_reduce needs backward computation.
I0413 06:08:56.442044 22206 net.cpp:220] inception_4b/relu_3x3 needs backward computation.
I0413 06:08:56.442047 22206 net.cpp:220] inception_4b/3x3 needs backward computation.
I0413 06:08:56.442050 22206 net.cpp:220] inception_4b/relu_3x3_reduce needs backward computation.
I0413 06:08:56.442054 22206 net.cpp:220] inception_4b/3x3_reduce needs backward computation.
I0413 06:08:56.442059 22206 net.cpp:220] inception_4b/relu_1x1 needs backward computation.
I0413 06:08:56.442062 22206 net.cpp:220] inception_4b/1x1 needs backward computation.
I0413 06:08:56.442066 22206 net.cpp:220] inception_4a/output_inception_4a/output_0_split needs backward computation.
I0413 06:08:56.442070 22206 net.cpp:220] inception_4a/output needs backward computation.
I0413 06:08:56.442075 22206 net.cpp:220] inception_4a/relu_pool_proj needs backward computation.
I0413 06:08:56.442078 22206 net.cpp:220] inception_4a/pool_proj needs backward computation.
I0413 06:08:56.442082 22206 net.cpp:220] inception_4a/pool needs backward computation.
I0413 06:08:56.442086 22206 net.cpp:220] inception_4a/relu_5x5 needs backward computation.
I0413 06:08:56.442090 22206 net.cpp:220] inception_4a/5x5 needs backward computation.
I0413 06:08:56.442095 22206 net.cpp:220] inception_4a/relu_5x5_reduce needs backward computation.
I0413 06:08:56.442097 22206 net.cpp:220] inception_4a/5x5_reduce needs backward computation.
I0413 06:08:56.442101 22206 net.cpp:220] inception_4a/relu_3x3 needs backward computation.
I0413 06:08:56.442106 22206 net.cpp:220] inception_4a/3x3 needs backward computation.
I0413 06:08:56.442109 22206 net.cpp:220] inception_4a/relu_3x3_reduce needs backward computation.
I0413 06:08:56.442112 22206 net.cpp:220] inception_4a/3x3_reduce needs backward computation.
I0413 06:08:56.442116 22206 net.cpp:220] inception_4a/relu_1x1 needs backward computation.
I0413 06:08:56.442121 22206 net.cpp:220] inception_4a/1x1 needs backward computation.
I0413 06:08:56.442124 22206 net.cpp:220] pool3/3x3_s2_pool3/3x3_s2_0_split needs backward computation.
I0413 06:08:56.442128 22206 net.cpp:220] pool3/3x3_s2 needs backward computation.
I0413 06:08:56.442132 22206 net.cpp:220] inception_3b/output needs backward computation.
I0413 06:08:56.442138 22206 net.cpp:220] inception_3b/relu_pool_proj needs backward computation.
I0413 06:08:56.442142 22206 net.cpp:220] inception_3b/pool_proj needs backward computation.
I0413 06:08:56.442147 22206 net.cpp:220] inception_3b/pool needs backward computation.
I0413 06:08:56.442150 22206 net.cpp:220] inception_3b/relu_5x5 needs backward computation.
I0413 06:08:56.442154 22206 net.cpp:220] inception_3b/5x5 needs backward computation.
I0413 06:08:56.442157 22206 net.cpp:220] inception_3b/relu_5x5_reduce needs backward computation.
I0413 06:08:56.442162 22206 net.cpp:220] inception_3b/5x5_reduce needs backward computation.
I0413 06:08:56.442165 22206 net.cpp:220] inception_3b/relu_3x3 needs backward computation.
I0413 06:08:56.442169 22206 net.cpp:220] inception_3b/3x3 needs backward computation.
I0413 06:08:56.442173 22206 net.cpp:220] inception_3b/relu_3x3_reduce needs backward computation.
I0413 06:08:56.442178 22206 net.cpp:220] inception_3b/3x3_reduce needs backward computation.
I0413 06:08:56.442181 22206 net.cpp:220] inception_3b/relu_1x1 needs backward computation.
I0413 06:08:56.442185 22206 net.cpp:220] inception_3b/1x1 needs backward computation.
I0413 06:08:56.442189 22206 net.cpp:220] inception_3a/output_inception_3a/output_0_split needs backward computation.
I0413 06:08:56.442199 22206 net.cpp:220] inception_3a/output needs backward computation.
I0413 06:08:56.442204 22206 net.cpp:220] inception_3a/relu_pool_proj needs backward computation.
I0413 06:08:56.442209 22206 net.cpp:220] inception_3a/pool_proj needs backward computation.
I0413 06:08:56.442212 22206 net.cpp:220] inception_3a/pool needs backward computation.
I0413 06:08:56.442216 22206 net.cpp:220] inception_3a/relu_5x5 needs backward computation.
I0413 06:08:56.442220 22206 net.cpp:220] inception_3a/5x5 needs backward computation.
I0413 06:08:56.442224 22206 net.cpp:220] inception_3a/relu_5x5_reduce needs backward computation.
I0413 06:08:56.442229 22206 net.cpp:220] inception_3a/5x5_reduce needs backward computation.
I0413 06:08:56.442232 22206 net.cpp:220] inception_3a/relu_3x3 needs backward computation.
I0413 06:08:56.442235 22206 net.cpp:220] inception_3a/3x3 needs backward computation.
I0413 06:08:56.442239 22206 net.cpp:220] inception_3a/relu_3x3_reduce needs backward computation.
I0413 06:08:56.442243 22206 net.cpp:220] inception_3a/3x3_reduce needs backward computation.
I0413 06:08:56.442247 22206 net.cpp:220] inception_3a/relu_1x1 needs backward computation.
I0413 06:08:56.442251 22206 net.cpp:220] inception_3a/1x1 needs backward computation.
I0413 06:08:56.442255 22206 net.cpp:220] pool2/3x3_s2_pool2/3x3_s2_0_split needs backward computation.
I0413 06:08:56.442260 22206 net.cpp:220] pool2/3x3_s2 needs backward computation.
I0413 06:08:56.442265 22206 net.cpp:220] conv2/norm2 needs backward computation.
I0413 06:08:56.442268 22206 net.cpp:220] conv2/relu_3x3 needs backward computation.
I0413 06:08:56.442272 22206 net.cpp:220] conv2/3x3 needs backward computation.
I0413 06:08:56.442276 22206 net.cpp:220] conv2/relu_3x3_reduce needs backward computation.
I0413 06:08:56.442281 22206 net.cpp:220] conv2/3x3_reduce needs backward computation.
I0413 06:08:56.442284 22206 net.cpp:220] pool1/norm1 needs backward computation.
I0413 06:08:56.442289 22206 net.cpp:220] pool1/3x3_s2 needs backward computation.
I0413 06:08:56.442293 22206 net.cpp:220] conv1/relu_7x7 needs backward computation.
I0413 06:08:56.442297 22206 net.cpp:220] conv1/7x7_s2 needs backward computation.
I0413 06:08:56.442301 22206 net.cpp:222] bb-obj-norm does not need backward computation.
I0413 06:08:56.442307 22206 net.cpp:222] bb-label-norm does not need backward computation.
I0413 06:08:56.442313 22206 net.cpp:222] obj-block_obj-block_0_split does not need backward computation.
I0413 06:08:56.442318 22206 net.cpp:222] obj-block does not need backward computation.
I0413 06:08:56.442325 22206 net.cpp:222] size-block_size-block_0_split does not need backward computation.
I0413 06:08:56.442329 22206 net.cpp:222] size-block does not need backward computation.
I0413 06:08:56.442335 22206 net.cpp:222] coverage-block does not need backward computation.
I0413 06:08:56.442342 22206 net.cpp:222] coverage-label_slice-label_4_split does not need backward computation.
I0413 06:08:56.442348 22206 net.cpp:222] obj-label_slice-label_3_split does not need backward computation.
I0413 06:08:56.442353 22206 net.cpp:222] size-label_slice-label_2_split does not need backward computation.
I0413 06:08:56.442358 22206 net.cpp:222] bbox-label_slice-label_1_split does not need backward computation.
I0413 06:08:56.442364 22206 net.cpp:222] foreground-label_slice-label_0_split does not need backward computation.
I0413 06:08:56.442370 22206 net.cpp:222] slice-label does not need backward computation.
I0413 06:08:56.442375 22206 net.cpp:222] val_transform does not need backward computation.
I0413 06:08:56.442380 22206 net.cpp:222] val_label does not need backward computation.
I0413 06:08:56.442384 22206 net.cpp:222] val_data does not need backward computation.
I0413 06:08:56.442387 22206 net.cpp:264] This network produces output loss_bbox
I0413 06:08:56.442391 22206 net.cpp:264] This network produces output loss_coverage
I0413 06:08:56.442395 22206 net.cpp:264] This network produces output mAP
I0413 06:08:56.442399 22206 net.cpp:264] This network produces output precision
I0413 06:08:56.442409 22206 net.cpp:264] This network produces output recall
I0413 06:08:56.442546 22206 net.cpp:284] Network initialization done.
I0413 06:08:56.443408 22206 solver.cpp:60] Solver scaffolding done.
I0413 06:08:56.448447 22206 caffe.cpp:231] Starting Optimization
I0413 06:08:56.448458 22206 solver.cpp:304] Solving
I0413 06:08:56.448462 22206 solver.cpp:305] Learning Rate Policy: step
I0413 06:08:56.454710 22206 solver.cpp:362] Iteration 0, Testing net (#0)
I0413 06:08:56.454725 22206 net.cpp:723] Ignoring source layer train_data
I0413 06:08:56.454730 22206 net.cpp:723] Ignoring source layer train_label
I0413 06:08:56.454733 22206 net.cpp:723] Ignoring source layer train_transform
I0413 06:09:23.007010 22206 solver.cpp:429] Test net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 06:09:23.007134 22206 solver.cpp:429] Test net output #1: loss_coverage = 305.735 (* 1 = 305.735 loss)
I0413 06:09:23.007150 22206 solver.cpp:429] Test net output #2: mAP = 0
I0413 06:09:23.007155 22206 solver.cpp:429] Test net output #3: precision = 0
I0413 06:09:23.007159 22206 solver.cpp:429] Test net output #4: recall = 0
I0413 06:09:40.952916 22206 solver.cpp:242] Iteration 0 (0 iter/s, 44.5051s/40 iter), loss = 317.739
I0413 06:09:40.952960 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 06:09:40.952968 22206 solver.cpp:261] Train net output #1: loss_coverage = 318.719 (* 1 = 318.719 loss)
I0413 06:09:40.952993 22206 sgd_solver.cpp:106] Iteration 0, lr = 0.001
I0413 06:12:09.236304 22206 solver.cpp:242] Iteration 40 (0.26975 iter/s, 148.286s/40 iter), loss = -6.43825e-20
I0413 06:12:09.236418 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 06:12:09.236426 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 06:12:09.236436 22206 sgd_solver.cpp:106] Iteration 40, lr = 0.001
I0413 06:14:37.587092 22206 solver.cpp:242] Iteration 80 (0.269627 iter/s, 148.353s/40 iter), loss = -6.43825e-20
I0413 06:14:37.587157 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 06:14:37.587165 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 06:14:37.587175 22206 sgd_solver.cpp:106] Iteration 80, lr = 0.001
I0413 06:17:05.949126 22206 solver.cpp:242] Iteration 120 (0.269607 iter/s, 148.364s/40 iter), loss = -6.43825e-20
I0413 06:17:05.949256 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 06:17:05.949266 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 06:17:05.949278 22206 sgd_solver.cpp:106] Iteration 120, lr = 0.001
I0413 06:19:34.343139 22206 solver.cpp:242] Iteration 160 (0.269549 iter/s, 148.396s/40 iter), loss = -6.43825e-20
I0413 06:19:34.343253 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 06:19:34.343263 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 06:19:34.343273 22206 sgd_solver.cpp:106] Iteration 160, lr = 0.001
I0413 06:22:02.687221 22206 solver.cpp:242] Iteration 200 (0.269639 iter/s, 148.346s/40 iter), loss = -6.43825e-20
I0413 06:22:02.687297 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 06:22:02.687307 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 06:22:02.687319 22206 sgd_solver.cpp:106] Iteration 200, lr = 0.001
I0413 06:24:31.148965 22206 solver.cpp:242] Iteration 240 (0.269426 iter/s, 148.464s/40 iter), loss = -6.43825e-20
I0413 06:24:31.149034 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 06:24:31.149042 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 06:24:31.149052 22206 sgd_solver.cpp:106] Iteration 240, lr = 0.001
I0413 06:26:59.616809 22206 solver.cpp:242] Iteration 280 (0.269415 iter/s, 148.47s/40 iter), loss = -6.43825e-20
I0413 06:26:59.616961 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 06:26:59.616971 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 06:26:59.616981 22206 sgd_solver.cpp:106] Iteration 280, lr = 0.001
I0413 06:29:27.940598 22206 solver.cpp:242] Iteration 320 (0.269676 iter/s, 148.326s/40 iter), loss = -6.43825e-20
I0413 06:29:27.940678 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 06:29:27.940688 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 06:29:27.940698 22206 sgd_solver.cpp:106] Iteration 320, lr = 0.001
I0413 06:29:31.628262 22206 solver.cpp:479] Snapshotting to binary proto file snapshot_iter_322.caffemodel
I0413 06:29:31.759322 22206 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_322.solverstate
I0413 06:31:57.534430 22206 solver.cpp:242] Iteration 360 (0.267387 iter/s, 149.596s/40 iter), loss = -6.43825e-20
I0413 06:31:57.534548 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 06:31:57.534557 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 06:31:57.534569 22206 sgd_solver.cpp:106] Iteration 360, lr = 0.001
I0413 06:34:26.713977 22206 solver.cpp:242] Iteration 400 (0.268129 iter/s, 149.182s/40 iter), loss = -6.43825e-20
I0413 06:34:26.714052 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 06:34:26.714061 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 06:34:26.714072 22206 sgd_solver.cpp:106] Iteration 400, lr = 0.001
I0413 06:36:56.036567 22206 solver.cpp:242] Iteration 440 (0.267872 iter/s, 149.325s/40 iter), loss = -6.43825e-20
I0413 06:36:56.036634 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 06:36:56.036643 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 06:36:56.036653 22206 sgd_solver.cpp:106] Iteration 440, lr = 0.001
I0413 06:39:25.477144 22206 solver.cpp:242] Iteration 480 (0.267661 iter/s, 149.443s/40 iter), loss = -6.43825e-20
I0413 06:39:25.477215 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 06:39:25.477223 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 06:39:25.477234 22206 sgd_solver.cpp:106] Iteration 480, lr = 0.001
I0413 06:41:54.831908 22206 solver.cpp:242] Iteration 520 (0.267815 iter/s, 149.357s/40 iter), loss = -6.43825e-20
I0413 06:41:54.831998 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 06:41:54.832008 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 06:41:54.832020 22206 sgd_solver.cpp:106] Iteration 520, lr = 0.001
I0413 06:44:24.388000 22206 solver.cpp:242] Iteration 560 (0.267454 iter/s, 149.558s/40 iter), loss = -6.43825e-20
I0413 06:44:24.388065 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 06:44:24.388074 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 06:44:24.388085 22206 sgd_solver.cpp:106] Iteration 560, lr = 0.001
I0413 06:46:53.393358 22206 solver.cpp:242] Iteration 600 (0.268443 iter/s, 149.008s/40 iter), loss = -6.43825e-20
I0413 06:46:53.393478 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 06:46:53.393488 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 06:46:53.393498 22206 sgd_solver.cpp:106] Iteration 600, lr = 0.001
I0413 06:49:22.734746 22206 solver.cpp:242] Iteration 640 (0.267839 iter/s, 149.344s/40 iter), loss = -6.43825e-20
I0413 06:49:22.734871 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 06:49:22.734881 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 06:49:22.734892 22206 sgd_solver.cpp:106] Iteration 640, lr = 0.001
I0413 06:49:33.901051 22206 solver.cpp:479] Snapshotting to binary proto file snapshot_iter_644.caffemodel
I0413 06:49:33.993005 22206 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_644.solverstate
I0413 06:51:52.247838 22206 solver.cpp:242] Iteration 680 (0.267531 iter/s, 149.515s/40 iter), loss = -6.43825e-20
I0413 06:51:52.248011 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 06:51:52.248023 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 06:51:52.248035 22206 sgd_solver.cpp:106] Iteration 680, lr = 0.001
I0413 06:54:21.442358 22206 solver.cpp:242] Iteration 720 (0.268103 iter/s, 149.197s/40 iter), loss = -6.43825e-20
I0413 06:54:21.442461 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 06:54:21.442471 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 06:54:21.442482 22206 sgd_solver.cpp:106] Iteration 720, lr = 0.001
I0413 06:56:50.671871 22206 solver.cpp:242] Iteration 760 (0.26804 iter/s, 149.232s/40 iter), loss = -6.43825e-20
I0413 06:56:50.671988 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 06:56:50.671998 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 06:56:50.672009 22206 sgd_solver.cpp:106] Iteration 760, lr = 0.001
I0413 06:59:19.819095 22206 solver.cpp:242] Iteration 800 (0.268188 iter/s, 149.149s/40 iter), loss = -6.43825e-20
I0413 06:59:19.819205 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 06:59:19.819213 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 06:59:19.819224 22206 sgd_solver.cpp:106] Iteration 800, lr = 0.001
I0413 07:01:49.744330 22206 solver.cpp:242] Iteration 840 (0.266796 iter/s, 149.927s/40 iter), loss = -6.43825e-20
I0413 07:01:49.744400 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 07:01:49.744408 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 07:01:49.744420 22206 sgd_solver.cpp:106] Iteration 840, lr = 0.001
I0413 07:04:18.976639 22206 solver.cpp:242] Iteration 880 (0.268035 iter/s, 149.234s/40 iter), loss = -6.43825e-20
I0413 07:04:18.976755 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 07:04:18.976765 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 07:04:18.976776 22206 sgd_solver.cpp:106] Iteration 880, lr = 0.001
I0413 07:06:48.090966 22206 solver.cpp:242] Iteration 920 (0.268247 iter/s, 149.116s/40 iter), loss = -6.43825e-20
I0413 07:06:48.091079 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 07:06:48.091089 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 07:06:48.091099 22206 sgd_solver.cpp:106] Iteration 920, lr = 0.001
I0413 07:09:17.116433 22206 solver.cpp:242] Iteration 960 (0.268407 iter/s, 149.028s/40 iter), loss = -6.43825e-20
I0413 07:09:17.116539 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 07:09:17.116549 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 07:09:17.116559 22206 sgd_solver.cpp:106] Iteration 960, lr = 0.001
I0413 07:09:35.753537 22206 solver.cpp:479] Snapshotting to binary proto file snapshot_iter_966.caffemodel
I0413 07:09:35.846139 22206 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_966.solverstate
I0413 07:11:46.418128 22206 solver.cpp:242] Iteration 1000 (0.26791 iter/s, 149.304s/40 iter), loss = -6.43825e-20
I0413 07:11:46.418980 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 07:11:46.418990 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 07:11:46.419003 22206 sgd_solver.cpp:106] Iteration 1000, lr = 0.001
I0413 07:14:15.607887 22206 solver.cpp:242] Iteration 1040 (0.268112 iter/s, 149.191s/40 iter), loss = -6.43825e-20
I0413 07:14:15.608037 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 07:14:15.608047 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 07:14:15.608058 22206 sgd_solver.cpp:106] Iteration 1040, lr = 0.001
I0413 07:16:44.633111 22206 solver.cpp:242] Iteration 1080 (0.268407 iter/s, 149.027s/40 iter), loss = -6.43825e-20
I0413 07:16:44.633185 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 07:16:44.633195 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 07:16:44.633205 22206 sgd_solver.cpp:106] Iteration 1080, lr = 0.001
I0413 07:19:13.658380 22206 solver.cpp:242] Iteration 1120 (0.268407 iter/s, 149.027s/40 iter), loss = -6.43825e-20
I0413 07:19:13.658449 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 07:19:13.658458 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 07:19:13.658468 22206 sgd_solver.cpp:106] Iteration 1120, lr = 0.001
I0413 07:21:42.794762 22206 solver.cpp:242] Iteration 1160 (0.268207 iter/s, 149.139s/40 iter), loss = -6.43825e-20
I0413 07:21:42.794872 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 07:21:42.794881 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 07:21:42.794893 22206 sgd_solver.cpp:106] Iteration 1160, lr = 0.001
I0413 07:24:12.001484 22206 solver.cpp:242] Iteration 1200 (0.268081 iter/s, 149.209s/40 iter), loss = -6.43825e-20
I0413 07:24:12.001596 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 07:24:12.001606 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 07:24:12.001617 22206 sgd_solver.cpp:106] Iteration 1200, lr = 0.001
I0413 07:26:41.062177 22206 solver.cpp:242] Iteration 1240 (0.268343 iter/s, 149.063s/40 iter), loss = -6.43825e-20
I0413 07:26:41.062252 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 07:26:41.062261 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 07:26:41.062273 22206 sgd_solver.cpp:106] Iteration 1240, lr = 0.001
I0413 07:29:10.076431 22206 solver.cpp:242] Iteration 1280 (0.268427 iter/s, 149.016s/40 iter), loss = -6.43825e-20
I0413 07:29:10.076552 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 07:29:10.076562 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 07:29:10.076573 22206 sgd_solver.cpp:106] Iteration 1280, lr = 0.001
I0413 07:29:36.118050 22206 solver.cpp:479] Snapshotting to binary proto file snapshot_iter_1288.caffemodel
I0413 07:29:36.210352 22206 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_1288.solverstate
I0413 07:31:39.303977 22206 solver.cpp:242] Iteration 1320 (0.268043 iter/s, 149.23s/40 iter), loss = -6.43825e-20
I0413 07:31:39.304044 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 07:31:39.304054 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 07:31:39.304064 22206 sgd_solver.cpp:106] Iteration 1320, lr = 0.001
I0413 07:34:10.779850 22206 solver.cpp:242] Iteration 1360 (0.264065 iter/s, 151.478s/40 iter), loss = -6.43825e-20
I0413 07:34:10.779914 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 07:34:10.779924 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 07:34:10.779934 22206 sgd_solver.cpp:106] Iteration 1360, lr = 0.001
I0413 07:36:40.017544 22206 solver.cpp:242] Iteration 1400 (0.268025 iter/s, 149.24s/40 iter), loss = -6.43825e-20
I0413 07:36:40.017660 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 07:36:40.017670 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 07:36:40.017683 22206 sgd_solver.cpp:106] Iteration 1400, lr = 0.001
I0413 07:39:09.247287 22206 solver.cpp:242] Iteration 1440 (0.268039 iter/s, 149.232s/40 iter), loss = -6.43825e-20
I0413 07:39:09.247429 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 07:39:09.247440 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 07:39:09.247452 22206 sgd_solver.cpp:106] Iteration 1440, lr = 0.001
I0413 07:41:38.391238 22206 solver.cpp:242] Iteration 1480 (0.268194 iter/s, 149.146s/40 iter), loss = -6.43825e-20
I0413 07:41:38.391366 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 07:41:38.391376 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 07:41:38.391387 22206 sgd_solver.cpp:106] Iteration 1480, lr = 0.001
I0413 07:44:07.598134 22206 solver.cpp:242] Iteration 1520 (0.26808 iter/s, 149.209s/40 iter), loss = -6.43825e-20
I0413 07:44:07.599515 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 07:44:07.599524 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 07:44:07.599535 22206 sgd_solver.cpp:106] Iteration 1520, lr = 0.001
I0413 07:46:36.945183 22206 solver.cpp:242] Iteration 1560 (0.267831 iter/s, 149.348s/40 iter), loss = -6.43825e-20
I0413 07:46:36.945298 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 07:46:36.945309 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 07:46:36.945320 22206 sgd_solver.cpp:106] Iteration 1560, lr = 0.001
I0413 07:49:06.170992 22206 solver.cpp:242] Iteration 1600 (0.268046 iter/s, 149.228s/40 iter), loss = -6.43825e-20
I0413 07:49:06.171119 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 07:49:06.171129 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 07:49:06.171140 22206 sgd_solver.cpp:106] Iteration 1600, lr = 0.001
I0413 07:49:39.681165 22206 solver.cpp:479] Snapshotting to binary proto file snapshot_iter_1610.caffemodel
I0413 07:49:39.773512 22206 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_1610.solverstate
I0413 07:49:39.850608 22206 solver.cpp:362] Iteration 1610, Testing net (#0)
I0413 07:49:39.850631 22206 net.cpp:723] Ignoring source layer train_data
I0413 07:49:39.850636 22206 net.cpp:723] Ignoring source layer train_label
I0413 07:49:39.850641 22206 net.cpp:723] Ignoring source layer train_transform
I0413 07:49:58.498661 22206 solver.cpp:429] Test net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 07:49:58.498684 22206 solver.cpp:429] Test net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 07:49:58.498710 22206 solver.cpp:429] Test net output #2: mAP = 0
I0413 07:49:58.498716 22206 solver.cpp:429] Test net output #3: precision = 0
I0413 07:49:58.498721 22206 solver.cpp:429] Test net output #4: recall = 0
I0413 07:51:54.017457 22206 solver.cpp:242] Iteration 1640 (0.23831 iter/s, 167.849s/40 iter), loss = -6.43825e-20
I0413 07:51:54.017525 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 07:51:54.017534 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 07:51:54.017545 22206 sgd_solver.cpp:106] Iteration 1640, lr = 0.001
I0413 07:54:23.078280 22206 solver.cpp:242] Iteration 1680 (0.268343 iter/s, 149.063s/40 iter), loss = -6.43825e-20
I0413 07:54:23.078408 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 07:54:23.078419 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 07:54:23.078431 22206 sgd_solver.cpp:106] Iteration 1680, lr = 0.001
I0413 07:56:52.164952 22206 solver.cpp:242] Iteration 1720 (0.268297 iter/s, 149.089s/40 iter), loss = -6.43825e-20
I0413 07:56:52.165076 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 07:56:52.165087 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 07:56:52.165098 22206 sgd_solver.cpp:106] Iteration 1720, lr = 0.001
I0413 07:59:21.380599 22206 solver.cpp:242] Iteration 1760 (0.268065 iter/s, 149.218s/40 iter), loss = -6.43825e-20
I0413 07:59:21.380702 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 07:59:21.380712 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 07:59:21.380722 22206 sgd_solver.cpp:106] Iteration 1760, lr = 0.001
I0413 08:01:50.508353 22206 solver.cpp:242] Iteration 1800 (0.268223 iter/s, 149.13s/40 iter), loss = -6.43825e-20
I0413 08:01:50.508491 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 08:01:50.508500 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 08:01:50.508513 22206 sgd_solver.cpp:106] Iteration 1800, lr = 0.001
I0413 08:04:19.601210 22206 solver.cpp:242] Iteration 1840 (0.268285 iter/s, 149.095s/40 iter), loss = -6.43825e-20
I0413 08:04:19.601328 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 08:04:19.601339 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 08:04:19.601351 22206 sgd_solver.cpp:106] Iteration 1840, lr = 0.001
I0413 08:06:48.672987 22206 solver.cpp:242] Iteration 1880 (0.268323 iter/s, 149.074s/40 iter), loss = -6.43825e-20
I0413 08:06:48.673053 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 08:06:48.673063 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 08:06:48.673074 22206 sgd_solver.cpp:106] Iteration 1880, lr = 0.001
I0413 08:09:17.874629 22206 solver.cpp:242] Iteration 1920 (0.26809 iter/s, 149.204s/40 iter), loss = -6.43825e-20
I0413 08:09:17.874747 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 08:09:17.874758 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 08:09:17.874768 22206 sgd_solver.cpp:106] Iteration 1920, lr = 0.001
I0413 08:09:58.770267 22206 solver.cpp:479] Snapshotting to binary proto file snapshot_iter_1932.caffemodel
I0413 08:09:58.862397 22206 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_1932.solverstate
I0413 08:11:46.843391 22206 solver.cpp:242] Iteration 1960 (0.268509 iter/s, 148.971s/40 iter), loss = -6.43825e-20
I0413 08:11:46.843502 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 08:11:46.843513 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 08:11:46.843523 22206 sgd_solver.cpp:106] Iteration 1960, lr = 0.001
I0413 08:14:15.769006 22206 solver.cpp:242] Iteration 2000 (0.268587 iter/s, 148.928s/40 iter), loss = -6.43825e-20
I0413 08:14:15.769112 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 08:14:15.769122 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 08:14:15.769134 22206 sgd_solver.cpp:106] Iteration 2000, lr = 0.001
I0413 08:16:44.733536 22206 solver.cpp:242] Iteration 2040 (0.268516 iter/s, 148.967s/40 iter), loss = -6.43825e-20
I0413 08:16:44.733660 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 08:16:44.733671 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 08:16:44.733681 22206 sgd_solver.cpp:106] Iteration 2040, lr = 0.001
I0413 08:19:13.831907 22206 solver.cpp:242] Iteration 2080 (0.268275 iter/s, 149.1s/40 iter), loss = -6.43825e-20
I0413 08:19:13.832031 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 08:19:13.832041 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 08:19:13.832053 22206 sgd_solver.cpp:106] Iteration 2080, lr = 0.001
I0413 08:21:42.788041 22206 solver.cpp:242] Iteration 2120 (0.268532 iter/s, 148.958s/40 iter), loss = -6.43825e-20
I0413 08:21:42.788151 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 08:21:42.788161 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 08:21:42.788172 22206 sgd_solver.cpp:106] Iteration 2120, lr = 0.001
I0413 08:24:11.707514 22206 solver.cpp:242] Iteration 2160 (0.268598 iter/s, 148.922s/40 iter), loss = -6.43825e-20
I0413 08:24:11.707672 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 08:24:11.707684 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 08:24:11.707695 22206 sgd_solver.cpp:106] Iteration 2160, lr = 0.001
I0413 08:26:40.647996 22206 solver.cpp:242] Iteration 2200 (0.26856 iter/s, 148.943s/40 iter), loss = -6.43825e-20
I0413 08:26:40.648113 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 08:26:40.648124 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 08:26:40.648134 22206 sgd_solver.cpp:106] Iteration 2200, lr = 0.001
I0413 08:29:09.702431 22206 solver.cpp:242] Iteration 2240 (0.268355 iter/s, 149.057s/40 iter), loss = -6.43825e-20
I0413 08:29:09.702499 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 08:29:09.702508 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 08:29:09.702519 22206 sgd_solver.cpp:106] Iteration 2240, lr = 0.001
I0413 08:29:58.160796 22206 solver.cpp:479] Snapshotting to binary proto file snapshot_iter_2254.caffemodel
I0413 08:29:58.253464 22206 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_2254.solverstate
I0413 08:31:38.717820 22206 solver.cpp:242] Iteration 2280 (0.268425 iter/s, 149.018s/40 iter), loss = -6.43825e-20
I0413 08:31:38.717890 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 08:31:38.717898 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 08:31:38.717908 22206 sgd_solver.cpp:106] Iteration 2280, lr = 0.001
I0413 08:34:07.770282 22206 solver.cpp:242] Iteration 2320 (0.268358 iter/s, 149.055s/40 iter), loss = -6.43825e-20
I0413 08:34:07.770356 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 08:34:07.770365 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 08:34:07.770376 22206 sgd_solver.cpp:106] Iteration 2320, lr = 0.001
I0413 08:36:36.791759 22206 solver.cpp:242] Iteration 2360 (0.268414 iter/s, 149.024s/40 iter), loss = -6.43825e-20
I0413 08:36:36.791877 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 08:36:36.791887 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 08:36:36.791899 22206 sgd_solver.cpp:106] Iteration 2360, lr = 0.001
I0413 08:39:05.874894 22206 solver.cpp:242] Iteration 2400 (0.268303 iter/s, 149.085s/40 iter), loss = -6.43825e-20
I0413 08:39:05.875010 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 08:39:05.875021 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 08:39:05.875032 22206 sgd_solver.cpp:106] Iteration 2400, lr = 0.001
I0413 08:41:34.985160 22206 solver.cpp:242] Iteration 2440 (0.268254 iter/s, 149.112s/40 iter), loss = -6.43825e-20
I0413 08:41:34.985235 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 08:41:34.985244 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 08:41:34.985255 22206 sgd_solver.cpp:106] Iteration 2440, lr = 0.001
I0413 08:44:03.968014 22206 solver.cpp:242] Iteration 2480 (0.268483 iter/s, 148.985s/40 iter), loss = -6.43825e-20
I0413 08:44:03.968082 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 08:44:03.968092 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 08:44:03.968103 22206 sgd_solver.cpp:106] Iteration 2480, lr = 0.001
I0413 08:46:33.126567 22206 solver.cpp:242] Iteration 2520 (0.268167 iter/s, 149.161s/40 iter), loss = -6.43825e-20
I0413 08:46:33.126674 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 08:46:33.126684 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 08:46:33.126695 22206 sgd_solver.cpp:106] Iteration 2520, lr = 0.001
I0413 08:49:02.207464 22206 solver.cpp:242] Iteration 2560 (0.268307 iter/s, 149.083s/40 iter), loss = -6.43825e-20
I0413 08:49:02.207599 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 08:49:02.207610 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 08:49:02.207622 22206 sgd_solver.cpp:106] Iteration 2560, lr = 0.001
I0413 08:49:58.178736 22206 solver.cpp:479] Snapshotting to binary proto file snapshot_iter_2576.caffemodel
I0413 08:49:58.270462 22206 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_2576.solverstate
I0413 08:51:31.441325 22206 solver.cpp:242] Iteration 2600 (0.268032 iter/s, 149.236s/40 iter), loss = -6.43825e-20
I0413 08:51:31.441433 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 08:51:31.441444 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 08:51:31.441455 22206 sgd_solver.cpp:106] Iteration 2600, lr = 0.001
I0413 08:54:00.436144 22206 solver.cpp:242] Iteration 2640 (0.268462 iter/s, 148.997s/40 iter), loss = -6.43825e-20
I0413 08:54:00.436259 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 08:54:00.436269 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 08:54:00.436280 22206 sgd_solver.cpp:106] Iteration 2640, lr = 0.001
I0413 08:56:29.580874 22206 solver.cpp:242] Iteration 2680 (0.268192 iter/s, 149.147s/40 iter), loss = -6.43825e-20
I0413 08:56:29.580984 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 08:56:29.580994 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 08:56:29.581006 22206 sgd_solver.cpp:106] Iteration 2680, lr = 0.001
I0413 08:58:58.719501 22206 solver.cpp:242] Iteration 2720 (0.268203 iter/s, 149.141s/40 iter), loss = -6.43825e-20
I0413 08:58:58.719606 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 08:58:58.719616 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 08:58:58.719629 22206 sgd_solver.cpp:106] Iteration 2720, lr = 0.001
I0413 09:01:27.701104 22206 solver.cpp:242] Iteration 2760 (0.268486 iter/s, 148.984s/40 iter), loss = -6.43825e-20
I0413 09:01:27.701212 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 09:01:27.701221 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 09:01:27.701232 22206 sgd_solver.cpp:106] Iteration 2760, lr = 0.001
I0413 09:03:56.649909 22206 solver.cpp:242] Iteration 2800 (0.268545 iter/s, 148.951s/40 iter), loss = -6.43825e-20
I0413 09:03:56.650017 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 09:03:56.650027 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 09:03:56.650038 22206 sgd_solver.cpp:106] Iteration 2800, lr = 0.001
I0413 09:06:25.543249 22206 solver.cpp:242] Iteration 2840 (0.268645 iter/s, 148.895s/40 iter), loss = -6.43825e-20
I0413 09:06:25.543360 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 09:06:25.543370 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 09:06:25.543381 22206 sgd_solver.cpp:106] Iteration 2840, lr = 0.001
I0413 09:08:54.533068 22206 solver.cpp:242] Iteration 2880 (0.268471 iter/s, 148.992s/40 iter), loss = -6.43825e-20
I0413 09:08:54.533181 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 09:08:54.533191 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 09:08:54.533202 22206 sgd_solver.cpp:106] Iteration 2880, lr = 0.001
I0413 09:09:57.946005 22206 solver.cpp:479] Snapshotting to binary proto file snapshot_iter_2898.caffemodel
I0413 09:09:58.037713 22206 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_2898.solverstate
I0413 09:11:23.805656 22206 solver.cpp:242] Iteration 2920 (0.267962 iter/s, 149.275s/40 iter), loss = -6.43825e-20
I0413 09:11:23.805763 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 09:11:23.805773 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 09:11:23.805783 22206 sgd_solver.cpp:106] Iteration 2920, lr = 0.001
I0413 09:13:52.782021 22206 solver.cpp:242] Iteration 2960 (0.268495 iter/s, 148.978s/40 iter), loss = -6.43825e-20
I0413 09:13:52.782095 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 09:13:52.782104 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 09:13:52.782114 22206 sgd_solver.cpp:106] Iteration 2960, lr = 0.001
I0413 09:16:21.811465 22206 solver.cpp:242] Iteration 3000 (0.268399 iter/s, 149.032s/40 iter), loss = -6.43825e-20
I0413 09:16:21.811580 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 09:16:21.811590 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 09:16:21.811601 22206 sgd_solver.cpp:106] Iteration 3000, lr = 0.001
I0413 09:18:50.943544 22206 solver.cpp:242] Iteration 3040 (0.268215 iter/s, 149.134s/40 iter), loss = -6.43825e-20
I0413 09:18:50.943651 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 09:18:50.943662 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 09:18:50.943673 22206 sgd_solver.cpp:106] Iteration 3040, lr = 0.001
I0413 09:21:20.150465 22206 solver.cpp:242] Iteration 3080 (0.26808 iter/s, 149.209s/40 iter), loss = -6.43825e-20
I0413 09:21:20.150588 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 09:21:20.150599 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 09:21:20.150609 22206 sgd_solver.cpp:106] Iteration 3080, lr = 0.001
I0413 09:23:49.286792 22206 solver.cpp:242] Iteration 3120 (0.268207 iter/s, 149.138s/40 iter), loss = -6.43825e-20
I0413 09:23:49.286911 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 09:23:49.286921 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 09:23:49.286932 22206 sgd_solver.cpp:106] Iteration 3120, lr = 0.001
I0413 09:26:18.334234 22206 solver.cpp:242] Iteration 3160 (0.268367 iter/s, 149.05s/40 iter), loss = -6.43825e-20
I0413 09:26:18.334349 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 09:26:18.334359 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 09:26:18.334370 22206 sgd_solver.cpp:106] Iteration 3160, lr = 0.001
I0413 09:28:47.519130 22206 solver.cpp:242] Iteration 3200 (0.26812 iter/s, 149.187s/40 iter), loss = -6.43825e-20
I0413 09:28:47.519201 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 09:28:47.519210 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 09:28:47.519222 22206 sgd_solver.cpp:106] Iteration 3200, lr = 0.0001
I0413 09:29:58.412418 22206 solver.cpp:479] Snapshotting to binary proto file snapshot_iter_3220.caffemodel
I0413 09:29:58.505291 22206 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_3220.solverstate
I0413 09:29:58.578584 22206 solver.cpp:362] Iteration 3220, Testing net (#0)
I0413 09:29:58.578608 22206 net.cpp:723] Ignoring source layer train_data
I0413 09:29:58.578613 22206 net.cpp:723] Ignoring source layer train_label
I0413 09:29:58.578616 22206 net.cpp:723] Ignoring source layer train_transform
I0413 09:30:17.187198 22206 solver.cpp:429] Test net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 09:30:17.187222 22206 solver.cpp:429] Test net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 09:30:17.187227 22206 solver.cpp:429] Test net output #2: mAP = 0
I0413 09:30:17.187232 22206 solver.cpp:429] Test net output #3: precision = 0
I0413 09:30:17.187237 22206 solver.cpp:429] Test net output #4: recall = 0
I0413 09:31:35.329571 22206 solver.cpp:242] Iteration 3240 (0.238361 iter/s, 167.813s/40 iter), loss = = 0 (* 1 = 0 loss)
I0413 16:03:01.155478 22206 sgd_solver.cpp:106] Iteration 9520, lr = 1e-05
I0413 16:05:31.172293 22206 solver.cpp:242] Iteration 9560 (0.266633 iter/s, 150.019s/40 iter), loss = -6.43825e-20
I0413 16:05:31.172405 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 16:05:31.172415 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 16:05:31.172428 22206 sgd_solver.cpp:106] Iteration 9560, lr = 1e-05
I0413 16:08:01.182947 22206 solver.cpp:242] Iteration 9600 (0.266644 iter/s, 150.013s/40 iter), loss = -6.43825e-20
I0413 16:08:01.183090 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 16:08:01.183101 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 16:08:01.183115 22206 sgd_solver.cpp:106] Iteration 9600, lr = 1e-06
I0413 16:10:31.031286 22206 solver.cpp:242] Iteration 9640 (0.266933 iter/s, 149.85s/40 iter), loss = -6.43825e-20
I0413 16:10:31.031422 22206 solver.cpp:261] Train net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 16:10:31.031432 22206 solver.cpp:261] Train net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 16:10:31.031445 22206 sgd_solver.cpp:106] Iteration 9640, lr = 1e-06
I0413 16:11:42.178346 22206 solver.cpp:479] Snapshotting to binary proto file snapshot_iter_9660.caffemodel
I0413 16:11:42.270309 22206 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_9660.solverstate
I0413 16:11:42.342959 22206 solver.cpp:362] Iteration 9660, Testing net (#0)
I0413 16:11:42.342983 22206 net.cpp:723] Ignoring source layer train_data
I0413 16:11:42.342988 22206 net.cpp:723] Ignoring source layer train_label
I0413 16:11:42.342991 22206 net.cpp:723] Ignoring source layer train_transform
I0413 16:12:01.052938 22206 solver.cpp:429] Test net output #0: loss_bbox = 0 (* 2 = 0 loss)
I0413 16:12:01.052963 22206 solver.cpp:429] Test net output #1: loss_coverage = 0 (* 1 = 0 loss)
I0413 16:12:01.052969 22206 solver.cpp:429] Test net output #2: mAP = 0
I0413 16:12:01.052974 22206 solver.cpp:429] Test net output #3: precision = 0
I0413 16:12:01.052979 22206 solver.cpp:429] Test net output #4: recall = 0
I0413 16:12:01.052984 22206 solver.cpp:347] Optimization Done.
I0413 16:12:01.052987 22206 caffe.cpp:234] Optimization Done.
What should i change ignorer to improve my accuracy from zero in DIGITS using caffe