Converting model from Caffe to ONNX - artificial-intelligence

Question

I'd like to prototype with Javascript to detect the sky using a model trained on a SkyFinder dataset.
I tried to convert the Caffe model (prototxt and trained data above) published here to the ONNX model using MMdnn.
Following command executed,
mmconvert --srcFramework caffe --inputWeight baseline.caffemodel --inputNetwork deploy.net --dstFramework onnx --outputModel baseline.onnx --inputShape 1 3 240 320
and this error came out
mmdnn.conversion.caffe.errors.ConversionError: Layer not found: label_0
Below is an excerpt from the first part of prototxt above. label_0 is connected to the dummy_data and convine_1 layers but does not appear to be defined anywhere.
name: "DeepBlueSky"
input: "data"
input_dim: 1
input_dim: 3
input_dim: 240
input_dim: 320
layers {
name: "dummy_data"
type: DUMMY_DATA
top: "label_0"
dummy_data_param {
data_filler {
type: "constant"
std: .5
}
num: 1
channels: 1
height: 240
width: 320
}
}
layers {
name: "combine_1"
type: CONCAT
bottom: "data"
bottom: "label_0"
top: "combine_1"
concat_param {
concat_dim: 1
}
}
...
I'm a newbie, so it's possible I'm missing the basics, but if you've had a similar experience and have been able to resolve it, would you please share it?