CAFFE2的MNIST的验证运行问题, 为何显示打不开LMDB数据库

 无论是从CAFFE2的安装目录 caffe2\caffe2\python\tutorials\ 中的找到的 MNIST.ipynb  还是从 官网 https://caffe2.ai/docs/tutorial-MNIST.html#main_wrap 下载的 代码, 均都是可以成功下载到lmdb文件 
lmdb train db found! lmdb test db found!
 training data folder:/home/xinwei/caffe2_notebooks/tutorial_data/mnist workspace root folder:/home/xinwei/caffe2_notebooks/tutorial_files/tutorial_mnist
 
 但是在juoyter notebook 中执行到workspace.RunNetOnce后,显示打不开数据库???

RuntimeError Traceback (most recent call last) <ipython-input-13-c329ccc61e68> in <module>() 1 # The parameter initialization network only needs to be run once. ----> 2 workspace.RunNetOnce(train_model.param_init_net) 3 # creating the network 4 workspace.CreateNet(train_model.net, overwrite=True) 5 # set the number of iterations and track the accuracy & loss /usr/local/caffe2/python/workspace.py in RunNetOnce(net) 208 C.Workspace.current._last_failed_op_net_position, 209 GetNetName(net), --> 210 StringifyProto(net), 211 ) 212 /usr/local/caffe2/python/workspace.py in CallWithExceptionIntercept(func, op_id_fetcher, net_name, *args, **kwargs) 191 def CallWithExceptionIntercept(func, op_id_fetcher, net_name, *args, **kwargs): 192 try: --> 193 return func(*args, **kwargs) 194 except Exception: 195 op_id = op_id_fetcher() 
 
RuntimeError: [enforce fail at db.h:206] db_. Cannot open db: /home/xinwei/caffe2_notebooks/tutorial_data/mnist/mnist-train-nchw-lmdb of type lmdb Error from operator: output: "dbreader_/home/xinwei/caffe2_notebooks/tutorial_data/mnist/mnist-train-nchw-lmdb" name: "" type: "CreateDB" arg { name: "db_type" s: "lmdb" } arg { name: "db" s: "/home/xinwei/caffe2_notebooks/tutorial_data/mnist/mnist-train-nchw-lmdb" }Now we can sample some of the data and predictions.
已邀请:

tensorflow2017

赞同来自:

下载的mnist测试验证代码没进行过任何的修订,下载lmdb正常,显示AI网络图也是正常的,但是到开始训练是就显示数据库打开异常,can't open
 
# The parameter initialization network only needs to be run once.
workspace.RunNetOnce(train_model.param_init_net)
# creating the network
workspace.CreateNet(train_model.net, overwrite=True)
# set the number of iterations and track the accuracy & loss
total_iters = 200
accuracy = np.zeros(total_iters)
loss = np.zeros(total_iters)
# Now, we will manually run the network for 200 iterations. 
for i in range(total_iters):
    workspace.RunNet(train_model.net)
    accuracy[i] = workspace.FetchBlob('accuracy')
    loss[i] = workspace.FetchBlob('loss')
# After the execution is done, let's plot the values.
pyplot.plot(loss, 'b')
pyplot.plot(accuracy, 'r')
pyplot.legend(('Loss', 'Accuracy'), loc='upper right')

lbd5788

赞同来自:

我也遇到这个问题,求教怎么解决

wywwork

赞同来自:

我也遇到了同样的问题,请问有谁解决了吗?

陶潜水

赞同来自:

我之前直接拿caffe生成的lmdb数据库在caffe2中训练测试就挂了,用caffe2中创建lmdb的脚本重新生成lmdb就ok,具体原因没有细究

要回复问题请先登录注册