load a trained model in other process with tflearn

Question

I am trying to save a model and load it in other process with tflearn library...

So I generated the model:

lenx = 21908
leny = 81
# Build neural network
net = tflearn.input_data(shape=[None, lenx])
net = tflearn.fully_connected(net, 8)
net = tflearn.fully_connected(net, 8)
net = tflearn.fully_connected(net, leny, activation='softmax')
net = tflearn.regression(net)

# Define model and setup tensorboard
model = tflearn.DNN(net, tensorboard_dir='tflearn_logs')
# Start training (apply gradient descent algorithm)
model.fit(train_x, train_y, n_epoch=10, batch_size=8, show_metric=True)
model.save('model.tflearn')

That works Ok! Then in other file, to run in other process I am trying to load it, in this way:

lenx = 21908
leny = 81
# Build neural network
net = tflearn.input_data(shape=[None, lenx])
net = tflearn.fully_connected(net, 8)
net = tflearn.fully_connected(net, 8)
net = tflearn.fully_connected(net, leny, activation='softmax')
net = tflearn.regression(net)

model = tflearn.DNN(net, tensorboard_dir='tflearn_logs')

model.load("model.tflearn")

But I got this error:

ValueError: Cannot feed value of shape (1, 0) for Tensor 'InputData/X:0', which has shape '(?, 21908)'

I have tried many things but it does not work.

Answer 1

The architecture of the neural net that you'll be loading should be the same. Also, you do not have to define the final layer, because there will be no training. Don't use the line net = tflearn.regression(net)

Answer 2

我认为您缺少的是负载中的“ weights_only = True”参数：

model.load("model.tflearn", weights_only=True)