Different results with the same parameter values ​​in the model(Random state)

[IpekInceler]

Hi Everyone,
First of all I like to use the Library. ❤️
I have a problem. I'm working on a time series data.
Although I use the same parameter values, I get different mape values ​​every time I run the model. Below is a screenshot of the parameters I use. I saw that there were no parameters such as random states or seeds that I could define in the model. How can I solve this problem??

Best regards and thanks,

Ipek

[ourownstory]

hi @IpekInceler and thank you!

This is documented at the end of Quickstart in chapter "Reproducibility"
https://ourownstory.github.io/neural_prophet/

For more details, have a look at #147

Let us know if this does not work for you!

[IpekInceler]

Hi @ourownstory
I looked at the link https://ourownstory.github.io/neural_prophet/ and solved my questions
Thanks so much for the suggestions.

[codeananda]

This is helpful, but I don't think it covers everything. The code for set_random_seed is just

np.random.seed(seed)
torch.manual_seed(seed)

This PyTorch thread discusses the many ways you can set a random seed and that the above two lines may not be enough. The most popular answer recommends doing this

    torch.manual_seed(seed)
    torch.cuda.manual_seed(seed)
    torch.cuda.manual_seed_all(seed)  # if you are using multi-GPU.
    np.random.seed(seed)  # Numpy module.
    random.seed(seed)  # Python random module.
    torch.manual_seed(seed)
    torch.backends.cudnn.benchmark = False
    torch.backends.cudnn.deterministic = True