I would be grateful if you could tell more about the use of other models: GPT-4, GPT-3.5, DALL·E and others. More information about the settings for each of the models: temperature, max_tokens and so on. If you can show more examples based on openai API, Google App Script and Telegret, that would be great! In any case, thanks!
Hi Ben, I have built the bot as explained in the video and your blog. However I am getting error with processUpdate function. When I deploy the script, bot keeps responding "please try again". Could you update the script and share us please? 😊
Thank you, and yes, but you will need to make a small tweak to how you pull the data out of the response object. When you use 'gpt-3.5-turbo', you pull the first response text like this: if (jsn.choices && jsn.choices.length > 0) { return jsn.choices[0].message.content; }
Is there a way to do this using WhatsApp? It seems that most of the work you did can be reused xcept the part about setting up the bot (using botfather). Do you know if there is a similar botfather service for WhatsApp?
Did you include a parameter in doPost, like this: function doPost(e) { ? If you did then there's something wrong with your Telegram webhook connection.
this is genius
Thanks for the videos and blog post with the code!
You're welcome! Let me know what other tutorials you'd like to see posted.
I would be grateful if you could tell more about the use of other models: GPT-4, GPT-3.5, DALL·E and others. More information about the settings for each of the models: temperature, max_tokens and so on. If you can show more examples based on openai API, Google App Script and Telegret, that would be great!
In any case, thanks!
Great idea; I'll look into that.
Very informative and innovative, thankyou so much
Thank you for sharing your feedback; it means a lot. Please let me know if there are other tutorials you'd like to see.
Thanks
Good tutorial, but how do I change the model to gpt-3.5-turbo?
You can see how I use the gpt-3.5-turbo in this video: th-cam.com/video/quvHnqHVzS0/w-d-xo.html at minute 28:30ish.
@@benronkin Thank u
Hi Ben, I have built the bot as explained in the video and your blog.
However I am getting error with processUpdate function. When I deploy the script, bot keeps responding "please try again".
Could you update the script and share us please? 😊
Hi, very nice code, is it possible to use gpt3 instead of davinci003 in this example, and how would code change?
Thank you, and yes, but you will need to make a small tweak to how you pull the data out of the response object. When you use 'gpt-3.5-turbo', you pull the first response text like this: if (jsn.choices && jsn.choices.length > 0) {
return jsn.choices[0].message.content;
}
@@benronkin thanks for reply
Is there a way to do this using WhatsApp? It seems that most of the work you did can be reused xcept the part about setting up the bot (using botfather). Do you know if there is a similar botfather service for WhatsApp?
That's a great idea. Let me look into that.
Error like this: TypeError: Cannot read properties of undefined (reading 'postData')
doPost., how to fix?
Did you include a parameter in doPost, like this: function doPost(e) { ? If you did then there's something wrong with your Telegram webhook connection.
@@benronkin am not sure my code, could you please send original code.sir
Sure, the code is available here: benronkin.com/blog/build-telegram-chat-bot-for-chatgpt-with-google-apps-script.html
@@benronkin let's me try, thank you sir.