Пікірлер
@tonyg_nerd
@tonyg_nerd Ай бұрын
I appreciate the vid, especially with JS. Anyone who creates a vid really needs to do their own walk-through once, before doing a vid, to avoid live issues. And when such an event occurs, edit.
@jinuvarghese9193
@jinuvarghese9193 4 ай бұрын
hi man, does sqltoolkit work with open ai agent?
@chang112x
@chang112x 4 ай бұрын
Thanks for the API key. 18:59
@codechit_en
@codechit_en 4 ай бұрын
😂 It won't work
@tk.gunshot.10
@tk.gunshot.10 6 ай бұрын
Please can you also do a video on how to use the openai api with langchain js using the updated method?
@tk.gunshot.10
@tk.gunshot.10 6 ай бұрын
Thank you for this. It was very helpful. Keep up the good work.
@sergiofitch4378
@sergiofitch4378 7 ай бұрын
This is a great explanation of the basics of using openAI's API. In the context of a chrome extension, how would you scale it for multiple users? You shouldn't expose the API to the client-side code, which, in chrome extensions is difficult since it is downloaded and installed on the client's browser.
@NIKHILGUPTA-if7uh
@NIKHILGUPTA-if7uh 8 ай бұрын
brother , i cant get that prompt which you are getting on right side of the window whenever you are searching something
@OstrichCode
@OstrichCode 9 ай бұрын
Thanks a lot. Can you do a video on integrating the Flowise Chain API as a custom tool, especially the one with multiple inputs?
@codechit_en
@codechit_en 9 ай бұрын
There is a sort of an error at the time of defining `temperature` field... It was spelling mistake. So, please correct that
@jagusiff
@jagusiff 10 ай бұрын
how can you tell whether the Conversational Retrieval QA Chain is using Knowledge from the Document vs. ChatGPT general knowledge when answering your prompt?
@codechit_en
@codechit_en 10 ай бұрын
There is a check, I forgot the name of it... But when you enable it then it will return you the piece of text from the document that it's referring to answer your question
@LaxmanKumar-g8t8o
@LaxmanKumar-g8t8o 10 ай бұрын
Thank you so much.
@its_yash0988
@its_yash0988 10 ай бұрын
I'm getting 401 (Too Many Requests) . how to correct ??
@MHA._19
@MHA._19 9 ай бұрын
+1
@MrNiceseb
@MrNiceseb 10 ай бұрын
Would vector embedding also be applicable if doing a NL2SQL where data is from SQL Server database? if so how to modify this flow
@codechit_en
@codechit_en 10 ай бұрын
Hi, this video was uploaded a few months ago, since then there might be some new update which I am not aware of, I would recommend you to check it's official documentation :)
@sachinibawanthi1765
@sachinibawanthi1765 10 ай бұрын
I'm getting 429 (Too Many Requests). But i didn't use my api key
@codechit_en
@codechit_en 10 ай бұрын
Create a new OpenAI key and make sure you haven't touched the limit
@PubgSpeed-wl8yo
@PubgSpeed-wl8yo 10 ай бұрын
The instructions are no longer up to date, the title cannot be underlined, Size cannot be added with it, and the number of tokens brought written without it does not work. Revise this instruction
@codechit_en
@codechit_en 10 ай бұрын
Thanks for feedback, I do understand that things have changed a lot since last I made this video, So, some of the parts might not be working, will update once I will get time 🙏
@biasedlyunbiased
@biasedlyunbiased 11 ай бұрын
Hello, I followed this but my bot seems clueless about the data. It's like it is not recognizing it. Any idea why?
@draxx8823
@draxx8823 11 ай бұрын
can you provide the github link for this repo
@KesavaRao-i8t
@KesavaRao-i8t Жыл бұрын
thanks for clear explanation
@pedromine308
@pedromine308 Жыл бұрын
Thanks for the video, how do you update the data ?
@pedromine308
@pedromine308 Жыл бұрын
can you use a dynamic variable for the pagesize ? I mean in the chat message change the number of the results ? thanks for the video and support :)
@PatrickCreations
@PatrickCreations Жыл бұрын
Does it still work up to today?
@codechit_en
@codechit_en Жыл бұрын
I doubt it, as they've updated an api
@AI-in-Action952
@AI-in-Action952 Жыл бұрын
how can we limit the Gpt model from answering unrelated to questions, i want to limit it only to my data
@codechit_en
@codechit_en Жыл бұрын
I think you can specify that in prompt strictly to do what and not or giving some examplar
@AI-in-Action952
@AI-in-Action952 Жыл бұрын
@@codechit_en Thanks a lot, but it's answering the general questions. shall i connect with in call for few mins
@tchisama
@tchisama Жыл бұрын
thanks for the help <3
@adimishra6950
@adimishra6950 Жыл бұрын
hii! atually the text completions are not showing up on the openAI websites do you know why is that?
@codechit_en
@codechit_en Жыл бұрын
Because they have replaced it with another model called gpt 3 instruct
@hernandocastroarana6206
@hernandocastroarana6206 Жыл бұрын
Thanks for the video. Do you know what components I can use in Flowise to make it throw citations at the end? Something similar to pihind or perplexity. I would like it to extract the information from websites or documents, but I want it to tell me where it extracted the information (citations) from. I would really appreciate the help
@codechit_en
@codechit_en Жыл бұрын
I think if you are using a vector database and embeddings then there is an option in conversational agents that gives you an option to `return source documents`
@malekaix
@malekaix Жыл бұрын
Hey again - hope all's well. Had a quick question if thats alright. So the API I'm calling via the custom tool is a Job Search api. Basically when I call that API, it returns LOTS of job info -- e.g. complete job descriptions for 10+ jobs. This returned data easily exceeds the OpenAI token limit, which causes my flow to error when I make the call. Is there a tool I could use to parse the returned Job Search data I give it to GPT? E.g. in a perfect world, I could parse the returned data and only provide GPT with the parsed company names and job titles. Thanks a ton man. Appreciate it!
@codechit_en
@codechit_en Жыл бұрын
Here you need little bit of programming knowledge, once you get the json of job search data, there would be lot of unnecessary token... To cater this issue you can create a string variable, and then loop through your result or json data... And at each iteration just extract company name and etc and append that inside variable... And finally return that text result
@malekaix
@malekaix Жыл бұрын
Oh that makes sense! I think I can ask GPT to update my code accordingly to make that happen. Thanks for this - really appreciate it@@codechit_en
@DilipKumar-uw3ly
@DilipKumar-uw3ly Жыл бұрын
@@malekaix - Will be great help, if you can share the solution once achieved. Thanks in advance.
@samayrathod7668
@samayrathod7668 Жыл бұрын
@samayrathod7668 0 seconds ago its showing error 429 again and again after i change my device and login with new user api key and run it then also it is showing 429 error Failed to load resource: the server responded with a status of 429 () content_scripts.js:62 Error: Error: Request failed with status: 429 at fetchCompletion (content_scripts.js:45:11)
@samayrathod7668
@samayrathod7668 Жыл бұрын
please help
@codechit_en
@codechit_en Жыл бұрын
It's an error by OpenAI when there are too many requests
@munyororo
@munyororo Жыл бұрын
Excellent work 🎉🙏🙌
@malekaix
@malekaix Жыл бұрын
thanks a lot for this!
@AIbel3arabi
@AIbel3arabi Жыл бұрын
Can you do a video on how to integrate calendly using custom tools?
@codechit_en
@codechit_en Жыл бұрын
Will try when I will get time
@educationforeveryone214
@educationforeveryone214 Жыл бұрын
Thanks you so much , Bro
@vava-b6w
@vava-b6w Жыл бұрын
Can we cite source URL after the response snippet?
@codechit_en
@codechit_en Жыл бұрын
I don't know if i get this correct, But if you are asking whether we can attatch url in response then yes we can because its jus a string, and whatever will be returned as response (string) will be passed as a context in openai to use
@vava-b6w
@vava-b6w Жыл бұрын
Thanks for the reply. I think it should not be difficult to hyperlink a title tag with URL string? would it be possible JS or HTML tag? What's your take on it?
@codechit_en
@codechit_en Жыл бұрын
Of course you can do that, if it's a string, because the only thing that tools care about is that you return strung
@RandomGuy4964
@RandomGuy4964 Жыл бұрын
awesome tutorial, exactly what I was looking for and the code actually works !
@samayrathod7668
@samayrathod7668 Жыл бұрын
its showing error 429 again and again after i change my device and login with new user api key and run it then also it is showing 429 error Failed to load resource: the server responded with a status of 429 () content_scripts.js:62 Error: Error: Request failed with status: 429 at fetchCompletion (content_scripts.js:45:11)
@lajeevkunalkripal4015
@lajeevkunalkripal4015 Жыл бұрын
@@samayrathod7668 same error is showing for me too
@ih4286
@ih4286 10 ай бұрын
@@samayrathod7668 I was getting that error yesterday it means '429 - Rate limit reached for requests Cause: You are sending requests too quickly.'
@temporary-user
@temporary-user Жыл бұрын
its not working
@codechit_en
@codechit_en Жыл бұрын
What's the problem, tell me the timestamp?
@YTSylantys
@YTSylantys 6 ай бұрын
if your getting an error about "429 too many requests" that means you need to add credit to openai api
@ajaybabupatel1665
@ajaybabupatel1665 Жыл бұрын
can you provide github link of this project
@klrah6291
@klrah6291 Жыл бұрын
brother can I put this in my resume.is it that level project. Will interviewer be impressed by this
@codechit_en
@codechit_en Жыл бұрын
I think you can, but i would recommend you to build something more complex and awesome than this... As it was just a tutorial to help people understand basic concepts:)
@vattory4277
@vattory4277 Жыл бұрын
thanks a lot
@codechit_en
@codechit_en Жыл бұрын
Snippet: quicknote.io/df431bb0-180e-11ee-89bb-0b5645e8e966
@sciencebabyy
@sciencebabyy Жыл бұрын
Yuu are the best
@gopinathkrm58
@gopinathkrm58 Жыл бұрын
how to restrict the response to a particular fields, if any question not related to specified topic then it should respond with 'not related' as response. Any idea on that ? Example I need only tech related responses, if user asks any non ques not related to tech, then it should always respond with 'not related'
@codechit_en
@codechit_en Жыл бұрын
There are two ways in which you can achieve something like this, although not sure as haven't tried anything like this before. First one is that you can have list of keywords related to technology like Tech, software, computer,etc... If given query doesn't contain any of these you can assume its not related to tech. 2. Pass user query to openai and ask if it related to technology if yes then return true or else return false But not actually sure though as haven't tried any of these :)
@gopinathkrm58
@gopinathkrm58 Жыл бұрын
@@codechit_en yes I tried that way, but sometimes it is not working, I mean it's answering some unwanted queries. I think I need to dig more into it. Thanks
@codechit_en
@codechit_en Жыл бұрын
Okay, and let me know if you find something interesting, as I am learning as well :)
@gopinathkrm58
@gopinathkrm58 Жыл бұрын
@@codechit_en okay
@tahmideditzps
@tahmideditzps Жыл бұрын
nice vid
@codechit_en
@codechit_en Жыл бұрын
One mistake that I made in this video was at 12:31 and something like that. where i directly put else part(In that case it will also run that code if val and node's val is equal). But in bst there shouldn't be any duplicate keys :)
@anigma3086
@anigma3086 Жыл бұрын
o bhai
@anigma3086
@anigma3086 Жыл бұрын
Bhagwa daari babua