Guys plan your queries before training the adapter. Once you type a query you can't remove it. Also you can't get more than 30 queries. Both cases you have to start from zero.
Thanks a lot for this tutorial video, it was very helpful! Is there a way to use these custom adapters in a python script, in which i pass an image an input and print the key value pairs in a JSON format?
Glad you enjoyed the video! 🙌 You may find this GitHub resource with Amazon Textract code examples helpful: go.aws/46K0npz. If not, feel free to ask our community of experts at re:Post for insight on this: go.aws/aws-repost. ✅🧐 ^MO
Thank you for the detailed video; it was very helpful. I have created an adapter with multiple custom queries, and now I want to use it in my .NET code to extract text from the documents. Can you help me with this? Thanks again.
Happy to hear the video was helpful! 😀This blog post has more info on customizing with Amazon Textract: go.aws/3zUAd70. 👈 Additionally, you can reach out to our experts over at re:Post for guidance: go.aws/aws-repost. 🧐 ^KS
This seems similar to AWS Comprehend NER? But much less half-baked, especially the annotating part. Can you explain why I would use one over the other?
I did some research, Amazon Textract is more like an OCR system. It can be used as a scanner for docs/images and it reads what's there and spits out the relevant text. Amazon Comprehend however, gathers insight about the text which it parsed. You can also do sentiment analysis e.g. When a text is extracted, Amazon Comprehend can give you insights like what are the nouns, positive or negative sentiment etc. Such insights are not possible with Textract.
Hi, there. This document discusses how you can use both Textract and Comprehend together, while including the sphere covered by each service: go.aws/3VwbPzS. These documents dig a bit deeper into the services: go.aws/3Vs8zFy & go.aws/3KKi3am. ℹ️ For more questions, please send them through to our free community of helpful specialists at re:Post: go.aws/aws-repost. ✍️ ^ZP
@@TommyWiggles I understand that. This custom query though is using ML to label the text that's extracted. It's quite similar to Comprehend's Entity Detection, at least for this example in the video where you're getting the check's amount, we can use comprehend entity detection to label the check amount. So I'm not sure when we use custom query vs comprehend's NER
Trying to call the adapter from code (python boto 3.12), but cant get it work. How is the Adapter response returned? Same as Queries json/structure? Do I have to specify the same queries from the adapter while calling textract.start_document_analysis() and use the FeatureTypes=["Queries"]? Seems redundant and the AWS documentation is not clear on that
Sorry to hear this. If you'd like, you can learn how to share feedback related to our docs, here: go.aws/documentation-feedback. 👈 We're always happy to hear how we can improve your experience. 📝 In the meantime, this is an excellent question to share with our community of experts over on re:Post: go.aws/aws-repost. 📬 ^SA
Hello! I found this doc to help you out: go.aws/3RUgeeZ. 📚 If this doesn't hit the mark, reach out to our community of experts and devs, in re:Post: go.aws/aws-repost. 🎯 ^RS
![เป็นไปได้ไหม - WanMai [Teaser] | 05-12-24](http://i.ytimg.com/vi/P5Uto6e453w/mqdefault.jpg)
damn u sound very serious! thanks for the tutorial, very much appreciated
We are glad it was helpful.👍
Finally, a video explaining in practice how it works. Thanks.
Guys plan your queries before training the adapter. Once you type a query you can't remove it. Also you can't get more than 30 queries. Both cases you have to start from zero.
Thanks a lot for this tutorial video, it was very helpful! Is there a way to use these custom adapters in a python script, in which i pass an image an input and print the key value pairs in a JSON format?
Glad you enjoyed the video! 🙌 You may find this GitHub resource with Amazon Textract code examples helpful: go.aws/46K0npz. If not, feel free to ask our community of experts at re:Post for insight on this: go.aws/aws-repost. ✅🧐 ^MO
Thank you for the detailed video; it was very helpful. I have created an adapter with multiple custom queries, and now I want to use it in my .NET code to extract text from the documents. Can you help me with this? Thanks again.
Happy to hear the video was helpful! 😀This blog post has more info on customizing with Amazon Textract: go.aws/3zUAd70. 👈 Additionally, you can reach out to our experts over at re:Post for guidance: go.aws/aws-repost. 🧐 ^KS
This seems similar to AWS Comprehend NER? But much less half-baked, especially the annotating part. Can you explain why I would use one over the other?
I did some research, Amazon Textract is more like an OCR system. It can be used as a scanner for docs/images and it reads what's there and spits out the relevant text. Amazon Comprehend however, gathers insight about the text which it parsed. You can also do sentiment analysis e.g. When a text is extracted, Amazon Comprehend can give you insights like what are the nouns, positive or negative sentiment etc. Such insights are not possible with Textract.
Hi, there. This document discusses how you can use both Textract and Comprehend together, while including the sphere covered by each service: go.aws/3VwbPzS. These documents dig a bit deeper into the services: go.aws/3Vs8zFy & go.aws/3KKi3am. ℹ️ For more questions, please send them through to our free community of helpful specialists at re:Post: go.aws/aws-repost. ✍️ ^ZP
@@TommyWiggles I understand that. This custom query though is using ML to label the text that's extracted. It's quite similar to Comprehend's Entity Detection, at least for this example in the video where you're getting the check's amount, we can use comprehend entity detection to label the check amount. So I'm not sure when we use custom query vs comprehend's NER
Trying to call the adapter from code (python boto 3.12), but cant get it work. How is the Adapter response returned? Same as Queries json/structure? Do I have to specify the same queries from the adapter while calling textract.start_document_analysis() and use the FeatureTypes=["Queries"]? Seems redundant and the AWS documentation is not clear on that
Sorry to hear this. If you'd like, you can learn how to share feedback related to our docs, here: go.aws/documentation-feedback. 👈 We're always happy to hear how we can improve your experience. 📝 In the meantime, this is an excellent question to share with our community of experts over on re:Post: go.aws/aws-repost. 📬 ^SA
How about line items? Do you know how can we identify them as array of Line Items?
Hello! I found this doc to help you out: go.aws/3RUgeeZ. 📚 If this doesn't hit the mark, reach out to our community of experts and devs, in re:Post: go.aws/aws-repost. 🎯 ^RS
@@awssupport Thank you! im going thru it now