You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I am trying to train a segmentation model from semantic-segmentation example on DIGITS. Training a Custom Network with the prototxt and FCN_Alexnet returns the following error:
ERROR: Cannot copy param 0 weights from layer 'fc6'; shape mismatch. Source param shape is 1 1 4096 9216 (37748736); target param shape is 4096 256 6 6 (37748736). To learn this layer's parameters from scratch rather than copying from a saved net, rename the layer.
norm1 needs backward computation.
pool1 needs backward computation.
relu1 needs backward computation.
conv1 needs backward computation.
shift does not need backward computation.
label_label_0_split does not need backward computation.
label does not need backward computation.
data_data_0_split does not need backward computation.
data does not need backward computation.
This network produces output accuracy
This network produces output loss
Network initialization done.
Solver scaffolding done.
Finetuning from /home/xhuv/DIGITS-5.0.0/fcn_alexnet.caffemodel
Attempting to upgrade input file specified using deprecated transformation parameters: /home/xhuv/DIGITS-5.0.0/fcn_alexnet.caffemodel
Successfully upgraded file specified using deprecated data transformation parameters.
Note that future Caffe releases will only support transform_param messages for transformation fields.
Attempting to upgrade input file specified using deprecated V1LayerParameter: /home/xhuv/DIGITS-5.0.0/fcn_alexnet.caffemodel
Successfully upgraded file specified using deprecated V1LayerParameter
Cannot copy param 0 weights from layer 'fc6'; shape mismatch. Source param shape is 1 1 4096 9216 (37748736); target param shape is 4096 256 6 6 (37748736). To learn this layer's parameters from scratch rather than copying from a saved net, rename the layer.
Should I change the prototxt with modifications on the layers?
The text was updated successfully, but these errors were encountered:
You'll want to change the name of the layer. I'm not sure we've documented this anywhere (it's really a Caffe thing, not a DIGITS thing), but I answered a similar SO post here: http://stackoverflow.com/a/39837047/2404152.
I am trying to train a segmentation model from semantic-segmentation example on DIGITS. Training a Custom Network with the prototxt and FCN_Alexnet returns the following error:
ERROR: Cannot copy param 0 weights from layer 'fc6'; shape mismatch. Source param shape is 1 1 4096 9216 (37748736); target param shape is 4096 256 6 6 (37748736). To learn this layer's parameters from scratch rather than copying from a saved net, rename the layer.
Should I change the prototxt with modifications on the layers?
The text was updated successfully, but these errors were encountered: