ProGamerGov

Here's the log file: https://gist.github.com/ProGamerGov/29219f98178a91ee3ddf039728db9bb3 `num_output` Increasing it by one means there is a new category composed of nothing, correct? Edit: Train.txt and Val.txt with labels: https://gist.github.com/ProGamerGov/6978038a0b40795289cafb554d9311af

This is from the model's research paper: > CNN feature. We use Caffe [28] for fine-tuning the > CNN model pre-trained on ImageNet [44]. **Images are resized > to 256...

Yes, it's CNN Object Proposal Models for Salient Object Detection which is a VGG16 model. https://github.com/BVLC/caffe/wiki/Model-Zoo > [VGG16:](https://gist.github.com/jimmie33/509111f8a00a9ece2c3d5dde6a750129) This model is used in the paper. > [GoogleNet:](https://gist.github.com/jimmie33/339fd0a938ed026692267a60b44c0c58) This model is...

Though the 256 LMDB results in `9] Check failed: datum_height == data_mean_.height() (256 vs. 224)` This is because your train_val.prototxt had `crop_size: 224`. But based on the research paper, it...

> There is some confusion here now. LMDB creation does not look into any prototxt, so I cannot understand how the LMDB creation could fail because of the crop in...

Maybe I just need to play around with the solver.prototxt some more to find the values that will let me rise above 31% accuracy. Or let it run for a...

> One thing to remember is that even if the training is not very successful, one can anyway always try how the model works in neural-style. Just trying iteration 5600...

I added these two layers into the train_val.prototxt to help understand how well training is going. ``` layers { name: "accuracy/top1" type: ACCURACY bottom: "fc8_43" bottom: "label" top: "accuracy@1" include:...

After no success in breaking past the 30-31% level of accuracy, I reinstalled everything on a fresh Ubuntu 16.04 with Cuda 8.0RC and Cudnn v5.

I recieved at error at iteration 8900: https://gist.github.com/ProGamerGov/4ac8b8ece45fd5a1a873636cdc673386 ``` I0731 07:23:18.979414 25048 solver.cpp:454] Snapshotting to binary proto file examples/imagenet/VGG16_SOD_finetune_from_scratch_iter_8900.caffemodel I0731 07:23:22.974056 25048 sgd_solver.cpp:273] Snapshotting solver state to binary proto file...