Questions tagged [google-cloud-tpu]

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.

Official website

188 questions
1
vote
1 answer

Where can I identify which Keras methods are TPU compatible methods?

I recently learned that UpSampling2D is not a TPU supported operation. How do I find out which Keras functions are compatible with a TPU in Colab? I was unable to find any documentation on the subject.
1
vote
1 answer

How can i converts the COCO dataset into a set of TFRecords on Google cloud TPU

I am trying to prepare the COCO dataset. on link https://cloud.google.com/tpu/docs/tutorials/retinanet#prepare_the_coco_dataset it is given "The tpu/tools/datasets/download_and_preprocess_coco.sh script converts the COCO dataset into a set of…
1
vote
1 answer

Google Cloud TPU support for Tensorflow C++ API

Does Google Cloud TPU support Tensorflow C++ API? I couldn't find any positive or negative statement about that, as well as a sample code or API reference. Does it mean it is not supported?
1
vote
1 answer

Save Model for Serving but "ValueError: Both labels and logits must be provided." when trying to export model

I wanted to save a model to do some predictions on specific pictures. Here is my serving function: def _serving_input_receiver_fn(): # Note: only handles one image at a time feat = tf.placeholder(tf.float32, shape=[None, 120, 50, 1]) …
craft
  • 495
  • 5
  • 16
1
vote
1 answer

Are there TPU compatible pre-trained checkpoints for image classification?

I have found TPU compatible pre-trained checkpoints for object detection but not for image classification. I'm specifically interested in Mobilenet_v2. I would switch to another network if such checkpoints were available and those for Mobilenet_v2…
jklieb
  • 41
  • 3
1
vote
2 answers

Starting Google Cloud TPU is missing VM image specification

I deleted my old cloud TPU instance and wanted to make a new one with ctpu up --preemptible. But now it expects a specification for the accompanying VM's image, since I get this error message: Created TPU proj-name! could not create Compute…
1
vote
1 answer

How to read logs before Deadline Exceeded on Init TPU system

I'm trying to run a model with Python 2.7 on a TPU with my own .tfrecord data file and all my code compiles, but the moment the TPU start doing its magic I don't have a clue what is going behind the scenes. Is there a way to track what is going on…
Victor
  • 504
  • 3
  • 9
1
vote
1 answer

Running Keras Functional API on TPU

I wanted to rewrite the create_model function into Keras Functional API. Running it on TPU though when I translate it gives me an error about using a Placeholder in create_method function. In the original example the author didn't put an explicit…
craft
  • 495
  • 5
  • 16
1
vote
2 answers

ModuleNotFoundError: No module named 'tensorflow.compiler'

~\AppData\Roaming\Python\Python36\site-packages\tensorflow\contrib\tpu\python\tpu\tpu_estimator.py in () 38 from tensorflow.contrib.tpu.python.tpu import tpu_config 39 from tensorflow.contrib.tpu.python.tpu import tpu_context --->…
1
vote
1 answer

Getting error in TPU with python mnist_tpu.py --use_tpu=false --tpu=''

I try to run this code from tensorflow git: mnist_tpu with python mnist_tpu.py --tpu='' python mnist_tpu.py --use_tpu=false --tpu='' but all the time I get the following error: AttributeError: module 'official.mnist.mnist' has no attribute…
craft
  • 495
  • 5
  • 16
1
vote
1 answer

Google cloud TPU: NotImplementedError: Non-resource Variables are not supported inside TPU computations

I am trying to train my model using google cloud's TPUs. The model works fine on CPU and GPU, and I can run the TPU tutorials without any problems (so it is not a problem of connecting to TPUs). However, when I run my program on the TPU cloud I get…
1
vote
2 answers

train_and_evaluate() batch size with TPU on GCMLE

I am trying to use TPUEstimator with train_and_evaluate() for an experiment on GCMLE. The TPUEstimator has a required argument train_batch_size that obviously specifies the batch size. However, for train_and_evaluate() I also specify a batch size…
1
vote
1 answer

Cloud TPU tools not generating profile

I've followed the instructions on Cloud TPU Tools. Except for step 4 where you have to change --tpu_name to --tpu, things seem to work as expected. What failed is the generation of the "Profile" tab. I executed capture_tpu_profile…
1
vote
1 answer

No OpKernel was registered to support Op 'ShutdownDistributedTPU' with these attrs. Registered devices

I'm trying to restore Mobile-net V2 model using TensorFlow 1.7.0 version from this link, and using the following code, but I am getting an error. import tensorflow as tf dir(tf.contrib) tf.reset_default_graph() v1 = tf.get_variable("v1",…
B Singh
  • 93
  • 1
  • 10
1
vote
1 answer

Train my model on GCP TPU

What architecture of program can run in TPU? Is it must to use tf.contrib.tpu.TPUEstimator? Is there any tpu-program example for reference except the tensorflow official model?