我尝试使用Caffe在Python中实现一个简单的丢失层是不成功的。作为参考,我发现在Python中实现了几个层,包括here,here和here。
从Caffe文档/示例提供的EuclideanLossLayer
开始,我无法使其正常工作并开始调试。即使使用这个简单的TestLayer
:
def setup(self, bottom, top):
"""
Checks the correct number of bottom inputs.
:param bottom: bottom inputs
:type bottom: [numpy.ndarray]
:param top: top outputs
:type top: [numpy.ndarray]
"""
print 'setup'
def reshape(self, bottom, top):
"""
Make sure all involved blobs have the right dimension.
:param bottom: bottom inputs
:type bottom: caffe._caffe.RawBlobVec
:param top: top outputs
:type top: caffe._caffe.RawBlobVec
"""
print 'reshape'
top[0].reshape(bottom[0].data.shape[0], bottom[0].data.shape[1], bottom[0].data.shape[2], bottom[0].data.shape[3])
def forward(self, bottom, top):
"""
Forward propagation.
:param bottom: bottom inputs
:type bottom: caffe._caffe.RawBlobVec
:param top: top outputs
:type top: caffe._caffe.RawBlobVec
"""
print 'forward'
top[0].data[...] = bottom[0].data
def backward(self, top, propagate_down, bottom):
"""
Backward pass.
:param bottom: bottom inputs
:type bottom: caffe._caffe.RawBlobVec
:param propagate_down:
:type propagate_down:
:param top: top outputs
:type top: caffe._caffe.RawBlobVec
"""
print 'backward'
bottom[0].diff[...] = top[0].diff[...]
我无法让Python层工作。学习任务相当简单,因为我只是想预测一个实数值是正数还是负数。相应的数据生成如下并写入LMDB:
N = 10000
N_train = int(0.8*N)
images = []
labels = []
for n in range(N):
image = (numpy.random.rand(1, 1, 1)*2 - 1).astype(numpy.float)
label = int(numpy.sign(image))
images.append(image)
labels.append(label)
将数据写入LMDB应该是正确的,因为使用Caffe提供的MNIST数据集的测试没有问题。网络定义如下:
net.data, net.labels = caffe.layers.Data(batch_size = batch_size, backend = caffe.params.Data.LMDB,
source = lmdb_path, ntop = 2)
net.fc1 = caffe.layers.Python(net.data, python_param = dict(module = 'tools.layers', layer = 'TestLayer'))
net.score = caffe.layers.TanH(net.fc1)
net.loss = caffe.layers.EuclideanLoss(net.score, net.labels)
使用以下方法手动完成解决:
for iteration in range(iterations):
solver.step(step)
相应的原型文件如下:
solver.prototxt
:
weight_decay: 0.0005
test_net: "tests/test.prototxt"
snapshot_prefix: "tests/snapshot_"
max_iter: 1000
stepsize: 1000
base_lr: 0.01
snapshot: 0
gamma: 0.01
solver_mode: CPU
train_net: "tests/train.prototxt"
test_iter: 0
test_initialization: false
lr_policy: "step"
momentum: 0.9
display: 100
test_interval: 100000
train.prototxt
:
layer {
name: "data"
type: "Data"
top: "data"
top: "labels"
data_param {
source: "tests/train_lmdb"
batch_size: 64
backend: LMDB
}
}
layer {
name: "fc1"
type: "Python"
bottom: "data"
top: "fc1"
python_param {
module: "tools.layers"
layer: "TestLayer"
}
}
layer {
name: "score"
type: "TanH"
bottom: "fc1"
top: "score"
}
layer {
name: "loss"
type: "EuclideanLoss"
bottom: "score"
bottom: "labels"
top: "loss"
}
test.prototxt
:
layer {
name: "data"
type: "Data"
top: "data"
top: "labels"
data_param {
source: "tests/test_lmdb"
batch_size: 64
backend: LMDB
}
}
layer {
name: "fc1"
type: "Python"
bottom: "data"
top: "fc1"
python_param {
module: "tools.layers"
layer: "TestLayer"
}
}
layer {
name: "score"
type: "TanH"
bottom: "fc1"
top: "score"
}
layer {
name: "loss"
type: "EuclideanLoss"
bottom: "score"
bottom: "labels"
top: "loss"
}
我尝试跟踪它,在backward
的{{1}}和foward
方法中添加调试消息,在解决过程中只调用TestLayer
方法(请注意,不测试执行时,调用只能与解决方案相关联)。同样,我在forward
中添加了调试消息:
python_layer.hpp
同样,只执行正向传递。当我删除virtual void Forward_cpu(const vector<Blob<Dtype>*>& bottom,
const vector<Blob<Dtype>*>& top) {
LOG(INFO) << "cpp forward";
self_.attr("forward")(bottom, top);
}
virtual void Backward_cpu(const vector<Blob<Dtype>*>& top,
const vector<bool>& propagate_down, const vector<Blob<Dtype>*>& bottom) {
LOG(INFO) << "cpp backward";
self_.attr("backward")(top, propagate_down, bottom);
}
中的backward
方法时,解决仍然有效。删除TestLayer
方法时,由于未实现forward
,因此会引发错误。我期望forward
也一样,所以似乎后向传递根本没有被执行。切换回常规图层并添加调试消息,一切都按预期工作。
我觉得我错过了一些简单或基本的东西,但我现在几天都无法解决问题。所以任何帮助或提示都会受到赞赏。
谢谢!
答案 0 :(得分:5)
这是预期的行为,因为您的python图层下方没有任何实际需要渐变来计算权重更新的图层。 Caffe注意到这一点并跳过这些层的反向计算,因为这会浪费时间。
如果在网络初始化时间日志中需要反向计算,则Caffe会打印所有图层。 在你的情况下,你应该看到类似的东西:
fc1 does not need backward computation.
如果在“Python”图层下面放置“InnerProduct”或“Convolution”图层(例如Data->InnerProduct->Python->Loss
),则需要进行反向计算,并调用后向方法。
答案 1 :(得分:2)
答案 2 :(得分:1)
即使我按照David Stutz的建议设置了force_backward: true
,我也没有工作。我发现here和here我忘记在目标类的索引处将最后一层的差异设置为1。
正如Mohit Jain在他的咖啡用户回答中所描述的那样,如果你正在用虎斑猫进行ImageNet分类,那么在做了前进传球之后,你将不得不这样做:
net.blobs['prob'].diff[0][281] = 1 # 281 is tabby cat. diff shape: (1, 1000)
请注意,您必须相应地将'prob'
更改为最后一个图层的名称,通常是softmax和'prob'
。
这是一个基于我的例子:
deploy.prototxt(它基于VGG16松散地显示文件的结构,但我没有测试它):
name: "smaller_vgg"
input: "data"
force_backward: true
input_dim: 1
input_dim: 3
input_dim: 224
input_dim: 224
layer {
name: "conv1_1"
type: "Convolution"
bottom: "data"
top: "conv1_1"
convolution_param {
num_output: 64
pad: 1
kernel_size: 3
}
}
layer {
name: "relu1_1"
type: "ReLU"
bottom: "conv1_1"
top: "conv1_1"
}
layer {
name: "pool1"
type: "Pooling"
bottom: "conv1_1"
top: "pool1"
pooling_param {
pool: MAX
kernel_size: 2
stride: 2
}
}
layer {
name: "fc1"
type: "InnerProduct"
bottom: "pool1"
top: "fc1"
inner_product_param {
num_output: 4096
}
}
layer {
name: "relu1"
type: "ReLU"
bottom: "fc1"
top: "fc1"
}
layer {
name: "drop1"
type: "Dropout"
bottom: "fc1"
top: "fc1"
dropout_param {
dropout_ratio: 0.5
}
}
layer {
name: "fc2"
type: "InnerProduct"
bottom: "fc1"
top: "fc2"
inner_product_param {
num_output: 1000
}
}
layer {
name: "prob"
type: "Softmax"
bottom: "fc2"
top: "prob"
}
main.py:
import caffe
prototxt = 'deploy.prototxt'
model_file = 'smaller_vgg.caffemodel'
net = caffe.Net(model_file, prototxt, caffe.TRAIN) # not sure if TEST works as well
image = cv2.imread('tabbycat.jpg', cv2.IMREAD_UNCHANGED)
net.blobs['data'].data[...] = image[np.newaxis, np.newaxis, :]
net.blobs['prob'].diff[0, 298] = 1
net.forward()
backout = net.backward()
# access grad from backout['data'] or net.blobs['data'].diff