2 Antworten
- Neueste
- Die meisten Stimmen
- Die meisten Kommentare
0
At the moment you can not custom train the Textrat Analzye Expense model. Try Textract Queries or the Comprehend route recommended by Indranil Banerjee AWS. Also file a ticket in your account to make the service team aware that you would like to add this field, so it can become part of the backlog.
beantwortet vor 2 Jahren
0
Not sure if this will solve your problem, but have you taken a look at this blog - https://aws.amazon.com/blogs/machine-learning/extracting-custom-entities-from-documents-with-amazon-textract-and-amazon-comprehend/
Relevanter Inhalt
AWS OFFICIALAktualisiert vor einem Jahr- AWS OFFICIALAktualisiert vor einem Jahr
- AWS OFFICIALAktualisiert vor einem Jahr
I'm not sure if its the same case. I think that Amazon Textract and Amazon Comprehend can works together when I want to train my custom model. But I want to train existing model creted by Amazon where I analyze Expenses - invoice. And this might be problem, because if my custom fields are not recognized firstly, there is no way how to get them later - because using Amazon Comprehend is service used After Amazon Textract. This is how I understand communication between these services.
In example you posted, they're probably extracting all the data with Amazon Textract and then specify what data should be in output with Amazon Comprehend.
The second way how to implement it is to use Amazon Textract to extract all the data and then only filter data I want. But it's not productive as there is way to use prebuilt models - invoices in Textract
Am I correct?