Prediciting the model with features on TPU (AssertionError: Could not compute output Tensor("dense_3/Identity:0", shape=(None, 1), dtype=float32))

[prudhvirajboddu]

Please make sure that this is a bug. As per our
GitHub Policy,
we only address code/doc bugs, performance issues, feature requests and
build/installation issues on GitHub. tag:bug_template

System information

• Have I written custom code (as opposed to using a stock example script provided in TensorFlow):
• OS Platform and Distribution (e.g., Linux Ubuntu 16.04): Kaggle using TPU
• Mobile device (e.g. iPhone 8, Pixel 2, Samsung Galaxy) if the issue happens on mobile device:
• TensorFlow installed from (source or binary): pip
• TensorFlow version (use command below): 2.2
• Python version: 3.7.6
• Bazel version (if compiling from source):
• GCC/Compiler version (if compiling from source):
• CUDA/cuDNN version:
• GPU model and memory:

You can collect some of this information using our environment capture
script
You can also obtain the TensorFlow version with:

1. TF 1.0: python -c "import tensorflow as tf; print(tf.GIT_VERSION, tf.VERSION)"
2. TF 2.0: python -c "import tensorflow as tf; print(tf.version.GIT_VERSION, tf.version.VERSION)"

Describe the current behavior
After succesful Training , prediction of the model was not happening even if the dataset was right
And throwing this error
AssertionError: Could not compute output Tensor("dense_3/Identity:0", shape=(None, 1), dtype=float32)
Describe the expected behavior
Model should sucessfully predict the dataset

Standalone code to reproduce the issue
Provide a reproducible test case that is the bare minimum necessary to generate
the problem. If possible, please share a link to Colab/Jupyter/any notebook.
https://www.kaggle.com/prudhvi9999/image-and-metadata

Other info / logs Include any logs or source code that would be helpful to
diagnose the problem. If including tracebacks, please include the full
traceback. Large logs and files should be attached.
Logs are
logs.txt

[amahendrakar]

@prudhvirajboddu,
I was able to run the code without any issues, please find the gist of it here. Thanks!

[prudhvirajboddu]

That was the latest version of changed code Not exactly I want the model to run.
This was the kernel version 2 . I was not able to run . Please take a look at it .
Thank you.
https://www.kaggle.com/prudhvi9999/image-and-metadata?scriptVersionId=36605669

[prudhvirajboddu]

That was the latest version of changed code Not exactly I want the model to run.
This was the kernel version 2 . I was not able to run . Please take a look at it .
Thank you.
https://www.kaggle.com/prudhvi9999/image-and-metadata?scriptVersionId=36605669

@amahendrakar any update ?

[hamzarabi3]

I'm having the same issue on Kaggle with TPU, model fit ran successfully, but I can't generate predictions.
AssertionError: Could not compute output Tensor("dense_7/Identity:0", shape=(None, 1), dtype=float32)
Temporary fix:
since mode.predict() can accept the format of training data and ignores the label, you can leverage this using map() on test data:
def prepare_test_data(sample):
----return sample, 0
just make sure to do it before batching.

[prudhvirajboddu]

I'm having the same issue on Kaggle with TPU, model fit ran successfully, but I can't generate predictions.
AssertionError: Could not compute output Tensor("dense_7/Identity:0", shape=(None, 1), dtype=float32)
Temporary fix:
since mode.predict() can accept the format of training data and ignores the label, you can leverage this using map() on test data:
def prepare_test_data(sample):
----return sample, 0
just make sure to do it before batching.

Can you share the code snippet for that ?

[hamzarabi3]

This worked for me
def prepare_test_data(sample):
----return sample, 0

test_ds = get_test_dataset()
test_images_ds = test_ds.map(prepare_test_data).batch(BATCH_SIZE)
ouput = model.predict(test_images_ds)