Caffe-Caffe Models

1.bvlc_reference_caffenet.caffemodel

下載地址：http://dl.caffe.berkeleyvision.org/bvlc_reference_caffenet.caffemodel

來源：caffe中ImageNet tutorial（Brew ImagesnNet）訓練結果

Paper：ImageNet Classification with Deep Convolutional Neural Networks（NIPS2012  Alex Krizhevsky   Ilya Sutskever   Geoffrey E. Hinton）

This model is the result of following the Caffe ImageNet model training instructions. It is a replication of the model described in the AlexNet publication with some differences:

• • not training with the relighting data-augmentation;
  • the order of pooling and normalization layers is switched (in CaffeNet, pooling is done before normalization).

This model is snapshot of iteration 310,000. The best validation performance during training was iteration 313,000 with validation accuracy 57.412% and loss 1.82328. This model obtains a top-1 accuracy 57.4% and a top-5 accuracy 80.4% on the validation set, using just the center crop. (Using the average of 10 crops, (4 + 1 center) * 2 mirror, should obtain a bit higher accuracy still.)

This model was trained by Jeff Donahue @jeffdonahue

AlexNet：

deploy.prototxt:

train_val.prototxt:

2.LeNet

3.finetune_flickr_style.caffemodel

Target：

        Fine-tuning CaffeNet for Style Recognition on “Flickr Style” Data.

Info：

        Fine-tuning takes an already learned model, adapts the architecture, and resumes training from the already learned model weights. Let’s fine-tune the BVLC-distributed CaffeNet model on a different dataset, Flickr Style, to predict image style instead of object category.

結構圖如下：

deploy.prototxt

train_val.prototxt