FiveTech Software tech support forums

... as cgi, for now very simple and simple, configured to receive requests and respond to requests. It reads the Request and accordingly generates the responses that are data from .dbf tables, which reside on the PC and at that point the prg program hard-coded opens tables, and generates responses ...

... new that the FiveWinners of this forum and the FiveWin Brasil forum are complaining about Mr. Silvio Falconi's attitude. Always with rude and rude responses, including with the Moderators of this wonderful forum. You have my full support to suspend it temporarily or permanently. Perfecto maestro ...

... and answers related to this topic and write the answer in json format" 2. You tell chatgpt to do it again without including previous responses Once we have the dataset then we train tinyLlama and we get our own specialized IA model :-)

Dear Antonio, thank you so much for taking your time to answer. I added the oWnd but now it seems that I no longer get all the responses as intended. here is what intentionally happens without oWnd: https://mybergland.com/ruth/webview2.jpg and may I also share the code without ...

... new that the FiveWinners of this forum and the FiveWin Brasil forum are complaining about Mr. Silvio Falconi's attitude. Always with rude and rude responses, including with the Moderators of this wonderful forum. You have my full support to suspend it temporarily or permanently. Perfecto maestro ...

Hi Antonio, Thank you. Question, if I set nSecs = 5 and responses comes back faster, will it still wait 5 seconds or process immediately? Does not wait. Processes immediately. Also for additional information: httprequest method SetTimeOuts( nResolveTimeOut, ...

Hi Antonio,
Thank you. Question, if I set nSecs = 5 and responses comes back faster, will it still wait 5 seconds or process immediately?

... to logical units and can then simply send these blocks to the AI and have it edited. I am particularly surprised by the rapid progress in the responses and the good results. Even harbour/FIVEWIN questions are now being answered well. A few months ago, there were practically no hits. Best regards, ...

... my statement was essentially, this response takes a lack of information, and from that draws erroneous conclusions, and tries to defend them with responses that do not actually relate to your question. AI is interesting, but it is sadly incomplete. Even so, people today tend to believe whatever ...

The problem with so much of those responses is they contain very generic statements. Clearly the input into the data used for those responses reflects a bias to "what is known" by the person who did the input. It reminds me of the ...

Dear Otto,

Here it properly creates the EXE and runs fine

> Possibly it's being deleted by my antivirus software right away?

Yes, probably

> Antonio, do you think one can also query the responses?

We need to find the Windows Copilot keys shortcut to copy the responses

... don't see where it creates the exe. Possibly it's being deleted by my antivirus software right away? Antonio, do you think one can also query the responses? I am using this approach with mod harbour but I add my prompts with copy and paste. The problem with the API key is that there is no cost ...

... Choose a Web Framework: Since Mod Harbour is primarily used for backend development, you'll need a web framework to handle HTTP requests and responses. Some popular choices include: Harbour Web - This is a built-in web server and HTTP library that comes with Harbour. CGI Programming - You ...

... support? But it is already possible now for you to include information in your requests, and ChatGPT can access that information when generating responses. I am extending my program. In a second memo field, I save knowledge bases, and then I can include them in my requests. Best regards, Otto

Hello Otto; I don't think you need to ask for number of tokens used. Responses contain that information already. Here is a typical Json response I get: { "id": "cmpl-XxXXXXXXXXXXXXXXXXX", "object": "text_completion", ...