Fine Tuning of GoogLeNet Model

by

Assuming you are trying to do image classification. These should be the steps for finetuning a model:

1. Classification layer

The original classification layer "loss3/classifier" outputs predictions for 1000 classes (it’s mum_output is set to 1000). You’ll need to replace it with a new layer with appropriate num_output. Replacing the classification layer:

  1. Change layer’s name (so that when you read the original weights from caffemodel file there will be no conflict with the weights of this layer).
  2. Change num_output to the right number of output classes you are trying to predict.
  3. Note that you need to change ALL classification layers. Usually there is only one, but GoogLeNet happens to have three: "loss1/classifier", "loss2/classifier" and "loss3/classifier".

2. Data

You need to make a new training dataset with the new labels you want to fine tune to. See, for example, this post on how to make an lmdb dataset.

3. How extensive a finetuning you want?

When finetuning a model, you can train ALL model’s weights or choose to fix some weights (usually filters of the lower/deeper layers) and train only the weights of the top-most layers. This choice is up to you and it ususally depends on the amount of training data available (the more examples you have the more weights you can afford to finetune).
Each layer (that holds trainable parameters) has param { lr_mult: XX }. This coefficient determines how susceptible these weights to SGD updates. Setting param { lr_mult: 0 } means you FIX the weights of this layer and they will not be changed during the training process.
Edit your train_val.prototxt accordingly.

4. Run caffe

Run caffe train but supply it with caffemodel weights as an initial weights:

~$ $CAFFE_ROOT/build/tools/caffe train -solver /path/to/solver.ptototxt -weights /path/to/orig_googlenet_weights.caffemodel

More Related Contents:

  1. Caffe | solver.prototxt values setting strategy
  2. How to create caffe.deploy from train.prototxt
  3. Why binary_crossentropy and categorical_crossentropy give different performances for the same problem?
  4. A guide to convert_imageset.cpp
  5. Common causes of nans during training of neural networks
  6. How to interpret caffe log with debug_info?
  7. Intuitive understanding of 1D, 2D, and 3D convolutions in convolutional neural networks [closed]
  8. What is `weight_decay` meta parameter in Caffe?
  9. Calculate the output size in convolution layer [closed]
  10. Caffe: What can I do if only a small batch fits into memory?
  11. How to reduce a fully-connected (`”InnerProduct”`) layer using truncated SVD
  12. What is `lr_policy` in Caffe?
  13. Is there anyone that can help me in finding an accurate dataset?
  14. What is a `”Python”` layer in caffe?
  15. Many to one and many to many LSTM examples in Keras
  16. How to verify CuDNN installation?
  17. Logo recognition in images [closed]
  18. How does Pytorch’s “Fold” and “Unfold” work?
  19. Epoch vs Iteration when training neural networks [closed]
  20. How to calculate optimal batch size
  21. What is the difference between loss function and metric in Keras? [duplicate]
  22. Keras Text Preprocessing – Saving Tokenizer object to file for scoring
  23. ResNet: 100% accuracy during training, but 33% prediction accuracy with the same data
  24. Does Caffe need data to be shuffled?
  25. Tackling Class Imbalance: scaling contribution to loss and sgd
  26. Is deep learning bad at fitting simple non linear functions outside training scope (extrapolating)?
  27. multi-layer perceptron (MLP) architecture: criteria for choosing number of hidden layers and size of the hidden layer? [closed]
  28. How to interpret loss and accuracy for a machine learning model [closed]
  29. Keras: model.evaluate vs model.predict accuracy difference in multi-class NLP task
  30. What is the difference between Keras model.evaluate() and model.predict()?

Leave a Comment