MetaEditor, Open AI and ChatGPT - page 5
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
The problem with AI is that people believe it will solve problems they themselves cannot solve. AI is not an intelligence in itself; it is an automation tool that learns what it is taught.
To use AI effectively, we must teach it to solve a problem in order to harness its capabilities and enhance its processes. AI will not create a magical solution unless someone has previously taught it to do so.
The problem with using LLMs to code in MQL is not a lack of creativity (they have that skill to some extent), or bad prompting, or inflated expectations. The problem is that the dataset for machine learning doesn't contain enough good quality data, so GTPs, for example, don't know the syntax, functions and architecture of MQL.
I feel that right now an experienced programmer like any of you who are already familiar with the framework could prompt the AI properly, but not a newbie. And with excessive prompting you're already doing to much work for it, while it's supposed to be doing things for you. Right now it's ironically quicker to manually code things yourself than deal with seriously fragile and revised prompting
Sergey , posts 43 and onwards are not related i'm afraid .
There is no any special summary thread about ONNX so I am using this summary thread to be the thread about ONNX as well/too.
So, as a result - there are some threads/posts about ONNX too:
--------------------------
If someone (who understands ONNX with all the particularities) starts the summary thread about ONNX (on the same way as Alain Verleyen started this summary thread) so it will be good, and some posts from here may be moved to new thread.
But for now - we are having this thread only sorry.
There is no any special summary thread about ONNX so I am using this summary thread to be the thread about ONNX as well/too.
So, as a result - there are some threads/posts about ONNX too:
--------------------------
If someone (who understands ONNX with all the particularities) starts the summary thread about ONNX (on the same way as Alain Verleyen started this summary thread) so it will be good, and some posts from here may be moved to new thread.
But for now - we are having this thread only sorry.
I understand onnx . This is extremely off topic . Just saying . It could go in the python section because that is the main utility (for mql5 coders) of onnx .
I understand onnx . This is extremely off topic . Just saying . It could go in the python section because that is the main utility (for mql5 coders) of onnx .
I moved the several posts to this summary thread: MetaTrader 5 Python User Group - the summary
Yes, you are right.
I moved the several posts to this summary thread: MetaTrader 5 Python User Group - the summary
excellent
I have a theory that ChatGPT doesn’t know MQL5 because there’s no a lot of resources for MQL5 online unlike popular programming languages like python and Js.
It tries to, but fails most of the time in using Syntax, it tends to mix up Mql4 and 5, and using "i" functions a lot like iClose, iOpen etc. C++ is closest to MQL ,again, only syntax is a bit different, i am also convinced ChatGpt has no clue on how to write MQL code successfully. Its ok for someone to get some ideas i guess, but i wouldn't use it to be honest, it will drop you in trouble, and you will be having a headache afterwards. I think that it is unable to learn as well, because as they said, it learns from data input, and when we think about it, how many millions of people have struggled, even myself, then i need to tell it that it is using the logic wrong, or there are syntax errors, or the functions are duplicated, the only reply it will give is "Sorry for the oversight, here is the corrected code" but then it continues on with everything except the corrected code. Yeah, i have seen this, tried this, it does not work, and i will not use it
Hi Alain, I'm all set up on the OpenAI side of things, got credits and an API key which I entered in the IDE options as shown in your post, but no joy.
I get the error message:
The API key is good. Where do I set the "organization"?
Hi Alain, I have configured everything on the OpenAI side, got the credits and the API key I entered in the IDE options as shown in your post, but nothing works.
I get an error message:
API key is appropriate. Where do I set the "organisation"?
Try to manually send a request to the API, for example via Python. There are probably examples of requests in this language.