I can look into a video about this, but I think if you're using GPT-4 model with indexed documents it provides citations automatically on the response object. I didn't include them in this video but that's a good topic to explore more.
Great walk through! I had an issue with authentication when setting up the Blob storage, timestamp 8:37. I submitted a ticket for help. I wonder if there are authentication parameters that need to be setup first before trying to connect azure search with Azure Studio?
Hello, were you able to get past the error? I am facing a similar error a little later at th-cam.com/video/OdpqciXDKjY/w-d-xo.html that says "We couldn't connect your data + Ensure managed identity is enabled for your service. Status: 400" Thanks in advance!
Suggestion: We have so many videos like this Can't you try showing how to get both video and text as answer for questions based on pdf document (which have mixed pics and text like websites)
Hey thanks, if I'm understanding correctly - you'd upload the docs into an Azure blob storage folder and then tell the search service to index that whole folder - that should make it available to the AI.
🙏tk u very much for the sharing. any chance there's a video . codebase for a chat-like interface / experience using similar azure open ai studio? tk u!
During Ingestion in progress at 8:41, it says "We couldn't connect your data Failed to create index". I have 19 PDF files that vary in size from 400MB to 500 MB. Any hint what maybe going on? Thanks for the content.
@@alexthecodewolf upon further testing, find out, that if I upload just one PDF out of all 19 PDFs I have, it works fine and creates the issue (mentioned in the last comment) only if I upload them all. Any suggestions?
@@alexthecodewolf I am facing a similar error at this time frame that says "We couldn't connect your data + Ensure managed identity is enabled for your service. Status: 400" I am pretty sure I have configured managed identity for all the services. Is there any debug diagnostic to look at? Thanks in advance!
isnt the stanard pricing tier the cost for deployment of the search engine for a month? From my experience, its not based on usage but on time and thats kinda pricy
Hey, thanks for watching. My understanding is that there are both pay-as-you-go based on tokens as well as reserved instances models. However the options and pricing vary depending on the model, whether you have to train it, your region, etc. The Azure pricing calculator is usually a reliable way to figure out the cost of what you want to do: azure.microsoft.com/en-us/pricing/calculator/ azure.microsoft.com/en-us/pricing/details/cognitive-services/openai-service/
@@Z4rnii Ooh sorry I misunderstood about the AI search vs OpenAI service. Yes, search can get pretty expensive :D But there is a free tier to mess around with that doesn't have an hourly charge, based on this page: azure.microsoft.com/en-us/pricing/details/search/
Excellent end to end video. When I ingest my data 2 very very small test files the 'Assistant Setup' sits for what seems like ages. I get bored, refresh the GUI and I get 'data imported'. When i then ask questions i get 'Validation error at #/dataSources/0/AzureCognitiveSearch/parameters/indexName: Field required' Perhaps there is an issue, perhaps the issue is my impatience!?!?!?! In the video you pause while ingestion happens - how long would you expect?
Thanks for the feedback! The data ingestion didn't take too long for me - if I remember correctly it was less than a minute. Your error seems to imply that the search index didn't get created or wasn't named properly - were you able to verify your index exists?
Some models such as GPT 4 are good at understanding code and structured data. They can rewrite code from one language to another but the quality/accuracy might not be 100%.
Hello, thanks for the video. I am currently working on a project and i am finding a big issue. I would like to know from your expertise to see if there is a possible solution. I have documents with sensitive data that i have to extract, however it seems that the new security features do not allow to extract names, dni, nif, etc. Some months ago this solution worked but now, not anymore. Do you know if there is any way to fix it? Thank you so much, i hope you or someone can help me.
Hey, thanks for watching. This might be due to your content filtering settings - I've seen similar issues and created a video about this topic th-cam.com/video/9g0T0Bd-U4s/w-d-xo.htmlsi=zVuqvS9SVe0lqCl8
Blazor just keeps getting better. It's on fire.
Nice tutorial. Can you make a video on how to work with citations when our data is based on documents?
I can look into a video about this, but I think if you're using GPT-4 model with indexed documents it provides citations automatically on the response object. I didn't include them in this video but that's a good topic to explore more.
Great walk through! I had an issue with authentication when setting up the Blob storage, timestamp 8:37. I submitted a ticket for help. I wonder if there are authentication parameters that need to be setup first before trying to connect azure search with Azure Studio?
Hello, were you able to get past the error? I am facing a similar error a little later at th-cam.com/video/OdpqciXDKjY/w-d-xo.html that says "We couldn't connect your data + Ensure managed identity is enabled for your service. Status: 400" Thanks in advance!
Another great tutorial :)
Cool application! Thanks for sharing this project!
Glad you like it!
Suggestion: We have so many videos like this
Can't you try showing how to get both video and text as answer for questions based on pdf document (which have mixed pics and text like websites)
Great man, I have a question, how can I upload multiple docs..and it will search based on all those documents that I have uploaded ?
Hey thanks, if I'm understanding correctly - you'd upload the docs into an Azure blob storage folder and then tell the search service to index that whole folder - that should make it available to the AI.
@@alexthecodewolf As Azure database administrator how is Azure openAI useful?
Can you help us with example of javascript or typescript
🙏tk u very much for the sharing. any chance there's a video . codebase for a chat-like interface / experience using similar azure open ai studio? tk u!
During Ingestion in progress at 8:41, it says "We couldn't connect your data Failed to create index". I have 19 PDF files that vary in size from 400MB to 500 MB. Any hint what maybe going on? Thanks for the content.
Hey, this sounds like an authentication issue - does it look like your storage account and connection to it were created successfully?
@@alexthecodewolf upon further testing, find out, that if I upload just one PDF out of all 19 PDFs I have, it works fine and creates the issue (mentioned in the last comment) only if I upload them all. Any suggestions?
@@alexthecodewolf I am facing a similar error at this time frame that says "We couldn't connect your data + Ensure managed identity is enabled for your service. Status: 400" I am pretty sure I have configured managed identity for all the services. Is there any debug diagnostic to look at? Thanks in advance!
Can we automate the process of adding data source in chat playground and train the model for that data using Python.
can you use python for the front end upload the videos for the same
Thank You Bro
isnt the stanard pricing tier the cost for deployment of the search engine for a month? From my experience, its not based on usage but on time and thats kinda pricy
Hey, thanks for watching. My understanding is that there are both pay-as-you-go based on tokens as well as reserved instances models. However the options and pricing vary depending on the model, whether you have to train it, your region, etc. The Azure pricing calculator is usually a reliable way to figure out the cost of what you want to do:
azure.microsoft.com/en-us/pricing/calculator/
azure.microsoft.com/en-us/pricing/details/cognitive-services/openai-service/
@@alexthecodewolf i was reffering to the azure ai search standard pricing, because oh that, im trying to use different search possibilities:)
@@Z4rnii Ooh sorry I misunderstood about the AI search vs OpenAI service. Yes, search can get pretty expensive :D But there is a free tier to mess around with that doesn't have an hourly charge, based on this page:
azure.microsoft.com/en-us/pricing/details/search/
Is OpenAI is available to costumers via an application form only? Meaning, only businesses?
The last time I checked you had to apply via the form still, but I believe personal accounts can also get approved pretty quick - mine was.
Excellent end to end video. When I ingest my data 2 very very small test files the 'Assistant Setup' sits for what seems like ages. I get bored, refresh the GUI and I get 'data imported'. When i then ask questions i get 'Validation error at #/dataSources/0/AzureCognitiveSearch/parameters/indexName: Field required'
Perhaps there is an issue, perhaps the issue is my impatience!?!?!?!
In the video you pause while ingestion happens - how long would you expect?
Thanks for the feedback! The data ingestion didn't take too long for me - if I remember correctly it was less than a minute. Your error seems to imply that the search index didn't get created or wasn't named properly - were you able to verify your index exists?
can it understand codes and possible do a rewrite and store it in blob
Some models such as GPT 4 are good at understanding code and structured data. They can rewrite code from one language to another but the quality/accuracy might not be 100%.
@@alexthecodewolf thanks for the answer can you possible point me to a tutorial assuming the data is already in blob
Hello, thanks for the video. I am currently working on a project and i am finding a big issue. I would like to know from your expertise to see if there is a possible solution. I have documents with sensitive data that i have to extract, however it seems that the new security features do not allow to extract names, dni, nif, etc. Some months ago this solution worked but now, not anymore. Do you know if there is any way to fix it?
Thank you so much, i hope you or someone can help me.
Hey, thanks for watching. This might be due to your content filtering settings - I've seen similar issues and created a video about this topic th-cam.com/video/9g0T0Bd-U4s/w-d-xo.htmlsi=zVuqvS9SVe0lqCl8
give us mario.txt file