Different tfds version on grader?

Course Q&A TensorFlow: Data and Deployment Specialization Data Pipelines with TensorFlow Data Services
1

Continuing the discussion from Error: TypeError("init() missing 2 required positional arguments: 'op' and 'message'",):

Stuck for 2 weeks on this, help anyone ?

I haven’t found a work-around for this, and I am not sure how to increase the grader verbosity to debug it in more detail

2

Hi! :wave:

Can you make sure that all tfds.load calls have data_dir parameter passed?

I was finally able to get this to work by specifying the data_dir parameter while using tfds.load()

3

Hi Yuri

I indicated that I had tried this unsuccessfully, tried these two:

splits, info = tfds.load(“cats_vs_dogs:4..”, split=splitnames, with_info=True, data_dir=f"{getcwd()}/data",

as_supervised=False)

splits, info = tfds.load(“cats_vs_dogs:4..”, split=splitnames, with_info=True, data_dir=‘data’,

as_supervised=True)

Note both do work with the web version, it’s the grader that refuses it, probably because it is not using the same version

Full ipynb attached fyi

Thanks

O

4

thank you for sharing the notebook; please, refrain from sharing the notebooks publicly in the future, use DMs for this

5

Sorry, I was using email, thinking I was talking to support

How can we get support/grader admin confirming or explaining what’s going on there ?

Thanks

6

Hi @Olivier_van_de_Wiel. Sorry, I’ve missed your message. Were you able to resolve the issue?