search

DrSleep / tensorflow-deeplab-resnet

DeepLab-ResNet rebuilt in TensorFlow
MIT License
1.25k stars 429 forks source link

"Convert.py" Error #73

Closed MyVanitar closed 7 years ago

MyVanitar commented 7 years ago

Hello,

The DeeplabV2 has two caffemodels. one named init.caffemodel and the second named train_iter_20000.caffemodel

I tried both but I get this error: 

Check failed: target_blobs.size() == source_layer.blobs_size() (5 vs. 3) Incompatible number of blobs for layer bn_conv1
*** Check failure stack trace: ***
Aborted (core dumped)

Please consider this, or upload the Tensorflow weights somewhere. Thanks.

DrSleep commented 7 years ago

The link to the converted weights is provided. Please read Readme carefully before asking questions/opening new issues.

As for your error, it might be because of your caffe version. On 24 Apr 2017 6:20 pm, "Hesam Moshiri" notifications@github.com wrote:

Hello,

The DeeplabV2 has two caffemodels. one named init.caffemodel and the second named train_iter_20000.caffemodel

I tried both but I get this error:

Check failed: target_blobs.size() == source_layer.blobs_size() (5 vs. 3) Incompatible number of blobs for layer bn_conv1 Check failure stack trace: Aborted (core dumped)

Please consider this, or upload the Tensorflow weights somewhere. Thanks.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/DrSleep/tensorflow-deeplab-resnet/issues/73, or mute the thread https://github.com/notifications/unsubscribe-auth/AHemmHu_mBmPzwqVSbRPngGQwv7CxTx8ks5rzL3DgaJpZM4NGUNv .

MyVanitar commented 7 years ago

Oh sorry yes, you are right. I just missed that download link

But anyway before closing this issue, may I ask you what is the actual difference in practice between the resnet_init weights and resnet?

DrSleep commented 7 years ago

Init is the model pretrained on COCO (not on Pascal). Resnet is the init model trained on Pascal On 24 Apr 2017 6:56 pm, "Hesam Moshiri" notifications@github.com wrote:

Hh sorry yes, you are right. I just missed that download link

But anyway before closing this issue, may I ask you what is the difference between the resnet_init weights resnet?

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/DrSleep/tensorflow-deeplab-resnet/issues/73#issuecomment-296717451, or mute the thread https://github.com/notifications/unsubscribe-auth/AHemmF5jBeRBtmQQp-iBmQ9w3UckpcU8ks5rzMY7gaJpZM4NGUNv .