Machine learning in trading: theory, models, practice and algo-trading - page 3235

[Renat Akhtyamov]

It's been on the website for ages, why are you bothering?

Thevirtual meeting of the ONNX community started with the introduction - Automated Trading Systems - MQL5

//Do you read nothing but this thread? ;)

[Aleksey Vyazmikin]

Aleksey Nikolayev #:
Google onnx.

Thanks.

Did I understand correctly that a number of standard functions are supported, but not self-written ones?

[Aleksey Nikolayev]

Aleksey Vyazmikin #:

Thank you.

Did I understand correctly that a number of standard functions are supported, but not self-written ones?

Google (or CHATGPTit) onnx from scratch. I'm a poor substitute for AI.

[Aleksey Vyazmikin]

Aleksey Nikolayev #:
Google (or CHATGRTit) onnx from scratch. I make a poor AI replacement.

ChatGPT:

"
ONNX (Open Neural Network Exchange) is a deep learning model exchange format that is designed to store and transfer models between different deep learning frameworks and tools. The main idea behind ONNX is to provide a common format for representing models, regardless of how they were created.

ONNX is not designed to store arbitrary custom functions or code. It is designed to represent models that can be expressed as a computational graph consisting of layers and operations supported by standard deep learning operations such as convolutions, pooling, activations, etc.

If you have a custom function that you would like to integrate into a deep learning model and store in ONNX format, you may have to implement the function using ONNX-supported operations and layers, or rewrite it as a computational graph. User code or functions written in a programming language should be represented as part of this graph using standard operations.

"

[Aleksey Nikolayev]

Aleksey Vyazmikin #:

ChatGPT:

"
ONNX (Open Neural Network Exchange) is a deep learning model exchange format that is designed to store and transfer models between different deep learning frameworks and tools. The main idea behind ONNX is to provide a common format for representing models, regardless of how they were created.

ONNX is not designed to store arbitrary custom functions or code. It is designed to represent models that can be expressed as a computational graph consisting of layers and operations supported by standard deep learning operations such as convolutions, pooling, activations, etc.

If you have a custom function that you would like to integrate into a deep learning model and store in ONNX format, you may have to implement the function using ONNX-supported operations and layers, or rewrite it as a computational graph. User code or functions written in a programming language should be represented as part of this graph using standard operations.

"

To convert an exponential moving average (EMA) into ONNX format, you can use the ONNX Python API

6

. First, you need to create an ONNX model using the ONNX API. Then, you can add the EMA calculation to the model. One way to calculate EMA in Python is to use the Pandas library

1

2 .

Once you have the EMA calculation implemented in Python, you can use the ONNX API to convert the Python code into an ONNX model. Here is an example Python code for calculating EMA using Pandas:

****************************

This is the beginning of AI's answer to your question about EMA. Once again, I urge you to switch from making human brains to AI.

[Aleksey Vyazmikin]

Aleksey Nikolayev #:

To convert an exponential moving average (EMA) into ONNX format, you can use the ONNX Python API

6

. First, you need to create an ONNX model using the ONNX API. Then, you can add the EMA calculation to the model. One way to calculate EMA in Python is to use the Pandas library

1

2 .

Once you have the EMA calculation implemented in Python, you can use the ONNX API to convert the Python code into an ONNX model. Here is an example Python code for calculating EMA using Pandas:

****************************

This is the beginning of AI's answer to your question about EMA. Once again, I urge you to switch from making human brains to AI.

Don't be clever - I am not interested in functions from the library, but in my own functions. And the AI gave me an answer that you can't do it without a tambourine.

If you don't want to answer, it's better to keep silent instead of choking on bile.

Super toxic branch.

[mytarmailS]

)))

He's beautiful

[Maxim Dmitrievsky]

Everything is encoded as graphs with matrix operators in their nodes. Nothing supernatural. Each model has its own separate parser into this format and back. There is a program on github that allows you to visually create or edit these graphs, based on netron.

[Aleksey Nikolayev]

Aleksey Vyazmikin #:

Don't be clever - I'm not interested in functions from the library, but in my own functions. And AI gave me an answer that it is impossible to do it without tambourine.

If you don't want to answer, you'd better keep silent instead of choking on bile.

Super toxic branch.

Those who want to do something are looking for opportunities, those who don't want to do it are looking for reasons.

You yourself are intoxicating the thread, trying to force its participants to "work together for your benefit" with your manipulations.

[Aleksey Nikolayev]

Nevertheless, the topic of creating custom pipelines and then converting them to ONNX is very interesting, important and well deserves one or even several articles on the forum.