caffe

问题 I just tried to train the provided Caffenet network with my own lmdb file. I changed the fully connected to convolutional layer of depth 4096 and custom frame_size. Here is the code: weight_param = dict(lr_mult=1, decay_mult=1) bias_param = dict(lr_mult=2, decay_mult=0) learned_param = [weight_param, bias_param] batch_size = 256 # 0 means non updating parameters frozen_param = [dict(lr_mult=0)] * 2 def conv_relu(bottom, ks, nout, stride=1, pad=0, group=1, param=learned_param, weight_filler

问题 I used 16-layers VGGnet to extract features from an image. It outputs a 4096-dimensional feature vector. However, I need a 1024-dimensional vector. How do I further reduce this 4096-vector into 1024-vector? Do I need to add a new layer on top of fc7 ? 回答1: Yes, you need to add another layer on top of fc7 . This is how your last few layers should be like layers { bottom: "fc7" top: "fc7" name: "relu7" type: RELU } layers { bottom: "fc7" top: "fc7" name: "drop7" type: DROPOUT dropout_param {

问题 I train my data set using caffe. I set (in slover.prototxt ): test_iter: 1000 test interval: 1000 max_iter: 450000 base_lr: 0.0001 lr_policy: "step" step_size: 100000 The test accuracy is around 0.02 and test loss is around 1.6 at the first test. Then the test accuracy increase and the test loss decrease every test. At iter 32000 the test accuracy is 1 and the test loss is 0.45. Then the accuracy decrease and the loss increase. I think the loss is too large when accuracy is 1. How do I know

问题 Following are the input layers of my fine-tuned model: layer { type: "HDF5Data" name: "data" top: "Meta" hdf5_data_param { source: "/path/to/train.txt" batch_size: 50 } include { phase: TRAIN } } layer { name: "data" type: "ImageData" top: "X" top: "Labels" include { phase: TRAIN } transform_param { mirror: true crop_size: 227 mean_file: "data/ilsvrc12/imagenet_mean.binaryproto" } image_data_param { source: "/path/to/train.txt" batch_size: 50 new_height: 256 new_width: 256 } } layer { type:

问题 I trained a model using the MNIST example architecture (but on my own set of 3 image classes) and have been trying to integrate it into the C++ example. I modified the MNIST architecture file to make it similar to the deploy.prototxt file for the C++ example (replacing the train and test layers with the input layer). Unfortunately, when I run the C++ program it gives me the following error: F0827 14:57:28.427697 25511 insert_splits.cpp:35] Unknown bottom blob 'label' (layer 'accuracy', bottom

问题 I have the train and label data as data.mat . (I have 200 training data with 6000 features and labels are (-1, +1) that have saved in data.mat). I am trying to convert my data (train and test) in hdf5 and run Caffe using: load input.mat hdf5write('my_data.h5', '/new_train_x', single( permute(reshape(new_train_x,[200, 6000, 1, 1]),[4:-1:1] ) )); hdf5write('my_data.h5', '/label_train', single( permute(reshape(label_train,[200, 1, 1, 1]), [4:-1:1] ) ) , 'WriteMode', 'append' ); hdf5write('my

问题 I train the CaffeNet (more precisely Cifar10 model for two classes classification) model. Now the model is ready for detection. For the model testing with a single image, I use test_predict_imagenet.cpp . I haven't tested how fast the code can run for 640 x 480 image. My target is I like to have 5~10 frame/sec is just nice for offline detection. I understand that I need to implement multi-size detection (i.e. something like we do in face detection, original image size is re-sized for

问题 I'm experimenting with LeNet network as a binary classifier (yes, no). The first and several last layers in the configuration file for testing is the following: layer { name: "data" type: "ImageData" top: "data" top: "label" include { phase: TEST } transform_param { scale: 0.00390625 } image_data_param { source: "examples/my_example/test_images_labels.txt" batch_size: 1 new_height: 128 new_width: 128 } } ... layer { name: "ip2" type: "InnerProduct" bottom: "ip1" top: "ip2" param { lr_mult: 1

问题 We all know this python code can create caffe networks: n = caffe.NetSpec() n.data, n.label = L.Data(batch_size=batch_size, backend=P.Data.LMDB, source=lmdb, transform_param=dict(scale=1. / 255), ntop=2) n.conv1 = L.Convolution(n.data, kernel_size=5, num_output=20, weight_filler=dict(type='xavier')) n.pool1 = L.Pooling(n.conv1, kernel_size=2, stride=2, pool=P.Pooling.MAX) The layer's name is on the right of n. for example:"n.data",this layer's name is "data". Write simple code If I want to

问题 Caffe is a deep learning framework that also includes a Matlab interface (Matcaffe). While Caffe supports multiple GPUs it seems that Matcaffe does not at the moment. Does anyone know of a workaround? 来源： https://stackoverflow.com/questions/33446612/caffe-multi-gpu-support-with-matlab-matcaffe