I am having an issue with the code loading the ‘imagenet’ weights from the module tf.keras.applications.MobileNetV2 with the mentioned error being “Exception: URL fetch failure on https://storage.googleapis.com/tensorflow/keras-applications/mobilenet_v2/mobilenet_v2_weights_tf_dim_ordering_tf_kernels_1.0_160.h5: None – [Errno 99] Cannot assign requested address”. I have tried many ways to load the weights from different modules but that might affect the grading system, and particularly, I haven’t found anyone addressing this error and a solution to it. I hope to receive some help.
You are the third student to report this issue today.
During the day, I have tested my copy of the notebook, and it’s about 50/50 whether it will be able to load the imagenet model on any given try.
I believe the issue is that the server with that model is overloaded, and your request to load the model is timing-out.
Course staff has been notifed, but there’s not much they can do about issues with outside resources.
I recommend you try again during hours when the server may be less busy.
Perhaps the course staff will be able to explore the issue further.
1 Like
Thanks for your response, I will take the assignment again.
I am having the same issue. Not able to run that block i.e. download to get the weights
1 Like
Course staff is investigating.
Monitor this thread for updates.
No additional reports are needed.
The model can be made available within the assignment workspace to avoid this.
See the weights parameter here which accepts the path to a locally saved file to load from.
>>> import tensorflow as tf
>>> model = tf.keras.applications.MobileNetV2(weights="./mobilenet_v2_weights_tf_dim_ordering_tf_kernels_1.0_160.h5")
>>>
Course staff is working on an update to the notebook for this.
1 Like
I’m closing this thread, because there are multiple threads on the same issue.
Please follow this thread for updates on this issue.
1 Like