Google Cloud TPUs (Tensor Processing Units) accelerate machine learning workloads developed using TensorFlow. This tag is used for questions about using the Google Cloud TPU service. Topics can range from the service user experience, issues with the trainer program written with Tensorflow, project quota issues, security, authentication, etc.
Questions tagged [google-cloud-tpu]
188 questions
0
votes
3 answers
Tensorflow Object Detection API w/ TPU Training - Display more granular Tensorboard plots
I've been following this tutorial on the Tensorflow Object Detection API, and I've successfully trained my own object detection model using Google's Cloud TPUs.
However, the problem is that on Tensorboard, the plots I'm seeing only have 2 data…
Horace Lee
- 146
- 6
0
votes
1 answer
Error about the ctpu command to build TPU testing environment
I have tried to follow the quickstart of Google TPU setup as the following link.
https://cloud.google.com/tpu/docs/quickstart
When I run "ctpu up" in the google cloud shell,
it shows the following error.
*2018/08/05 09:37:08 Created TPU…
Thomas Lau
- 121
- 1
0
votes
2 answers
TPU Based Tuning for CloudML
Are TPUs supported for distributed hyperparameter search? I'm using the tensor2tensor library, which supports CloudML for hyperparameter search, i.e., the following works for me to conduct hyperparameter search for a language model on…
alizaidi
- 61
- 4
0
votes
1 answer
TPUEstimator error -- AttributeError: module 'tensorflow.contrib.tpu.python.ops.tpu_ops' has no attribute 'cross_replica_sum'
I have written a tensorflow code using the TPUEstimator, but I am having problems running it in use_tpu=False mode. I would like to run it on my local computer to make sure that all the operations are TPU-compatible. The code works fine with the…
Adrien Doerig
- 86
- 10
0
votes
1 answer
Producing summary for gradients via cloud TPU's host_call_fn()?
My understanding is that host_call and host_call_fn() transfer stats from the TPU to the host. However, the instructions are not very clear on how to generate summary for anything non-scalar.
For example, I've tried to modify the official…
0
votes
1 answer
Why "tf.nn.bidirectional_dynamic_rnn" has a "Incompatible Operations" on TPU when TensorBoard check it?
I meet a problem when I check the "TPU Compatibility" of a bidirectional rnn. The TensorBoard tell me a reversal operation of sequence length vector is incompatible on TPU. I don't know why?
My simple code:
X_batch = np.array([
…
Martin_at_Coventry
- 23
- 5
0
votes
2 answers
Unable to delete Cloud TPU
I have a TPU node name 'demo-tpu' which I am try to delete, using the command gcloud beta compute tpus delete demo-tpu
This results in the following error message:
ERROR: (gcloud.beta.compute.tpus.delete) ABORTED: The resource is already in the…
Harshit Malik
- 11
- 1
0
votes
2 answers
Running TensorFlow trainer with Cloud ML Engine on TPU produces google.rpc.QuotaFailure
I have developed a TensorFlow model on Cloud ML Engine with scaleTier: BASIC.
Running its trainer experimentally on a GPU with scaleTier: BASIC_GPU works fine. But an attempt of running it on a TPU with scaleTier: BASIC_TPU produces this error…
Drux
- 11,992
- 13
- 66
- 116