predictions = sess.run(prob_tensor, {input_node: [bgr_image]})

[TanviShrivastav]

Azure documentation issue guidance

Thanks for opening an issue in the Azure technical documentation repository.

We use GitHub issues as the primary channel for customer and community feedback about the Azure documentation.

Creating an issue

We prefer that you create documentation feedback issues using the Feedback link on the published article - the feedback control on the doc page creates an issue that contains all the article details so you can focus on the feedback part.

You can also create a feedback issue here in the repo. If you do this, please make sure your issue lists:

• [ ] The relevant Azure service or technology.
• [ ] A link to the published documentation article that you have feedback about.
• [ ] Clear, specific feedback that the author can act on.

Pull requests and article contributions

If you know the change that is needed in an article, we encourage you to submit the changes directly using a pull request. If the change is large, or if you want to contribute an entire article, follow these guidelines:

• [ ] Don't surprise us with a big pull request or a pull request with a new article! Submit an issue that describes the details of the proposed large change or new article.
• [ ] Include the service or technology area.

We'll route the issue to the appropriate content team for review and discussion.

Tech support and product feedback

If you would like to contact Microsoft about other things, such as product feedback or tech support, please review these guidelines:

• If you need technical support using Azure, the paid and free support options are described here: https://azure.microsoft.com/support/options/.

• Each article in the Azure technical documentation contains a product feedback button - it's best to submit product feedback directly from a relevant article. Otherwise, you can submit product feedback for most Azure products in the following product feedback forum: https://feedback.azure.com/forums/34192--general-feedback.

---

Document Details

⚠ Do not edit this section. It is required for docs.microsoft.com ➟ GitHub issue linking.

• ID: 754dd7d5-dd8b-4a75-5d79-0ba621a847cd
• Version Independent ID: 3cca3535-1c16-2da1-46c6-421cbf808bc3
• Content: Run TensorFlow model in Python - Custom Vision Service - Azure Cognitive Services
• Content Source: articles/cognitive-services/Custom-Vision-Service/export-model-python.md
• Service: cognitive-services
• Product: unspecified
• GitHub Login: @areddish
• Microsoft Alias: areddish

[TanviShrivastav]

predictions = sess.run(prob_tensor, {input_node: [bgr_image]}) NameError: name 'bgr_image' is not defined Couldnt fix this. Please help and let me know where the bgr_image is defined?

[TravisCragg-MSFT]

@TanviShrivastav could you please post the document that you are referring to?

[TanviShrivastav]

https://docs.microsoft.com/en-us/azure/cognitive-services/custom-vision-service/export-model-python This one. The last portion, in predicting the image. Thank you.

[TravisCragg-MSFT]

@TanviShrivastav Thanks for the feedback! I have assigned the issue to the content author to evaluate and update as appropriate.

[ksujanpaul]

predictions = sess.run(prob_tensor, {input_node: [bgr_image]}) NameError: name 'bgr_image' is not defined

yeah it is the only error even i couldn't resolve and there is no alternative to change the variable name 'bgr_image'

fix this issue asap please

[tanvi2797]

Bgr_image is an undefined parameter. Replace it with augmented_image as that is the last modification that we have done on the input image. It works for me now.

[oderza]

I agree with @tanvi2797 , all you need to do is replace bgr_image with augmented_image and it works like a charm. predictions = sess.run(prob_tensor, {input_node: [augmented_image]})

[ksujanpaul]

yeah got it thank you @tanvi2797 @oderza for the response

[TravisCragg-MSFT]

@ksujanpaul We will now proceed to close this thread. If there are further questions regarding this matter, please tag me in your reply. We will gladly continue the discussion and we will reopen the issue.