I have a problem with the TPUs on the C4_W4_Lab_3_CelebA_GAN_Experiments notebook. First of all tensorflow is not installed by default whenusing TPU, so I downloaded it using !pip install tensorflow (I also tried with ==2.16.1 and 2.18). Than I get this error: TPUs not found in the cluster. Failed …

Hi. We’ve pushed a temporary fix. You can use the TPU by using the fallback runtime. Please relaunch the notebook from the classroom to see the instructions. We’ll monitor updates from Colab regarding using the TPU for newer runtimes. As of now, they haven’t updated their docs yet. For now, this sho…

are you logged into your account? it shows the kernel not registered? I didn’t encounter this issue. it also points on system configuration.

yes I am logged in my account. These are errors popping out. No OpKernel was registered to support Op ‘ConfigureDistributedTPU’ used by {{node ConfigureDistributedTPU}} with these attrs: [compilation_failure_closes_chips=false, is_global_init=false, enable_whole_mesh_compilations=false, embedding_c…

can I know two things, one are you running colab by the course provided or your own colab? Also is your colab connected to your Google drive. the reason I am asking this, because the error you encountered could be due to either missing tensorflow package or version mismatch. You can Also try to c…

Hi everyone! Thank you for reporting this! We will look into the problem and update you asap.

I’m running my own saved copy of the lab. I cleared the browsing history and cache, then tried again. However, nothing has changed. Thank you so much for your assistance.

so you are running locally? and did you encounter this issue in course provided colab?

Yes, I ran the code in the Colab notebook provided by the course itself and encountered the same issue.

can you check the tensorflow version and keras version it is running on once

[image] Deepti_Prasad: check the tensorflow version and keras version TensorFlow version: 2.18.0 Keras version: 3.8.0

Hi! We see the problem with the labs and are actively working on a fix. Will update again here in a few hours with the progress. Thank you for reporting, and sorry for the delay!

C4_W4_Lab3_CelebA Problem with TPU

Course Q&A TensorFlow: Advanced Techniques Specialization Generative Deep Learning with TensorFlow

chris.favila April 6, 2026, 10:29am 30

Hello again. To avoid notifying previous learners in this topic, let’s move the discussion to this new topic. Will close this one. Thank you.

1 Like

Topic		Replies	Views
Can't connect to TPU runtime in C4_W4_Lab_3_CelebA_GAN_Experiments Generative Deep Learning with TensorFlow week-module-4 , coursera-platform	1	14	December 24, 2025
None of the C2W4 labs run in Colab Custom and Distributed Training with TF week-module-4	5	50	May 23, 2025
Ungraded Lab: KeyError: 'COLAB_TPU_ADDR' Generative Deep Learning with TensorFlow week-module-4	4	713	June 21, 2024
Ungraded Lab: CelebA GAN Experiments C4_W4_Lab_3 AttributeError: 'DistributedDataset' object has on attribute '_variant_tensor' Generative Deep Learning with TensorFlow week-module-4	6	444	August 24, 2023
C3-W2-Eager Few Shot Object Detection error Advanced Computer Vision with TensorFlow week-module-2	11	634	February 9, 2022

C4_W4_Lab3_CelebA Problem with TPU

Related topics