samelhousseini / km-openai

MIT License
56 stars 97 forks source link

Unable to deploy - Fails validation error #10

Open sridhar21111976 opened 1 year ago

sridhar21111976 commented 1 year ago

Looks like the Responsible AI terms needs to be agreed (a check box when deploying cognitive resources through Azure GUI).. The template has no provision for the same and fails in validation with the below error

{"code":"InvalidTemplateDeployment","details":[{"code":"ResourceKindRequireAcceptTerms","message":"This subscription cannot create CognitiveServices until you agree to Responsible AI terms for this resource. You can agree to Responsible AI terms by creating a resource through the Azure Portal then trying again. For more detail go to https://aka.ms/csrainotice"}],"message":"The template deployment 'Microsoft.Template-20230530170353' is not valid according to the validation procedure. The tracking id is '089e0507-f000-4a9c-94b8-6a794f633ac2'. See inner errors for details."}

samelhousseini commented 1 year ago

Hello Sridhar,

Yes, this happens if this is your first time ever to deploy a Cognitive Services resource. So accept the agreement (checkbox), and you won't have this error the next time you deploy this solution.

Thanks

Samer


From: sridhar21111976 @.> Sent: Tuesday, May 30, 2023 11:06 AM To: samelhousseini/km-openai @.> Cc: Subscribed @.***> Subject: [samelhousseini/km-openai] Unable to deploy - Fails validation error (Issue #10)

Looks like the Responsible AI terms needs to be agreed (a check box when deploying cognitive resources through Azure GUI).. The template has no provision for the same and fails in validation with the below error

{"code":"InvalidTemplateDeployment","details":[{"code":"ResourceKindRequireAcceptTerms","message":"This subscription cannot create CognitiveServices until you agree to Responsible AI terms for this resource. You can agree to Responsible AI terms by creating a resource through the Azure Portal then trying again. For more detail go to https://aka.ms/csrainotice"}],"message":"The template deployment 'Microsoft.Template-20230530170353' is not valid according to the validation procedure. The tracking id is '089e0507-f000-4a9c-94b8-6a794f633ac2'. See inner errors for details."}

— Reply to this email directly, view it on GitHubhttps://github.com/samelhousseini/km-openai/issues/10, or unsubscribehttps://github.com/notifications/unsubscribe-auth/AZ2RAUWF55DYMLRDP4WVDJTXIWL53ANCNFSM6AAAAAAYTR47IU. You are receiving this because you are subscribed to this thread.Message ID: @.***>

sridhar21111976 commented 1 year ago

Thank you Samer, that worked and went past that validation. But failed to deploy App Service plan... I am able to deploy App service plan outside this template though. Would you know whats causing this..! Also does the resources need to be in the same region as the Azure Open AI resource...?

samelhousseini commented 1 year ago

Hello Sridhar,

I am not sure, because I never had any problem with the App Service Plan (or had any feedback regarding this). I am pretty sure that this a restriction on your subscription regarding creating App Service Plans, please check with your administrator.

Resources do not need to be in the same region as the OpenAI resrouce.

Thanks

Samer


From: sridhar21111976 @.> Sent: Wednesday, May 31, 2023 7:48 AM To: samelhousseini/km-openai @.> Cc: Samer El Housseini @.>; Comment @.> Subject: Re: [samelhousseini/km-openai] Unable to deploy - Fails validation error (Issue #10)

Thank you Samer, that worked and went past that validation. But failed to deploy App Service plan... I am able to deploy App service plan outside this template though. Would you know whats causing this..! Also does the resources need to be in the same region as the Azure Open AI resource...?

— Reply to this email directly, view it on GitHubhttps://github.com/samelhousseini/km-openai/issues/10#issuecomment-1569453017, or unsubscribehttps://github.com/notifications/unsubscribe-auth/AZ2RAUXJ6UOHSHIPJRJNCGTXI25Q7ANCNFSM6AAAAAAYTR47IU. You are receiving this because you commented.Message ID: @.***>

sridhar21111976 commented 1 year ago

Thanks Samer, I got it working by deploying in a different region... I am encountering one last error in summarisation as below..

Any suggestions around this

KeyError: 'Could not automatically map gpt-35-turbo to a tokeniser. Please use tiktok.get_encoding to explicitly get the tokeniser you expect.

sridhar21111976 commented 1 year ago

Hi Samer,

I found the above error is only coming when using Map_reduce in summarisation...
Also can you point out if the addition of a new document to the storage container will automatically trigger processing and reindexing...!?

samelhousseini commented 1 year ago

Hi Sridhar,

For summarization specifically, I think it had problems with Python 3.8, so maybe try 3.9 or later.

For the triggering of the ingestion, this is done in two parts:

  1. if you're talking about the raw documents (pdfs, word, ppts, txt files, etc..) this is triggered explicitly through Cognitive Services indexer APIs
  2. if you're talking about the json files in the processed folder in blob storage, then as soon as a new json file is uploaded, the rest is automatically triggered through events

Thanks

Samer


From: sridhar21111976 @.> Sent: Thursday, June 1, 2023 11:15 AM To: samelhousseini/km-openai @.> Cc: Samer El Housseini @.>; Comment @.> Subject: Re: [samelhousseini/km-openai] Unable to deploy - Fails validation error (Issue #10)

Hi Samer,

I found the above error is only coming when using Map_reduce in summarisation... Also can you point out if the addition of a new document to the storage container will automatically trigger processing and reindexing...!?

— Reply to this email directly, view it on GitHubhttps://github.com/samelhousseini/km-openai/issues/10#issuecomment-1571489149, or unsubscribehttps://github.com/notifications/unsubscribe-auth/AZ2RAURDW4P5VYPXB3D3F7TXJA6QXANCNFSM6AAAAAAYTR47IU. You are receiving this because you commented.Message ID: @.***>

sridhar21111976 commented 1 year ago

Thank you Samer. Have implemented the complete solution. Just need to try the web crawler.... 👍 A documentation update on the exact sequence of flow of data within the architecture will be super useful... Much appreciate the work done so far on this... Super useful.

sridhar21111976 commented 1 year ago

Thank you for supporting