Drastic performance drop

[hectorhernandezbabbel]

Hi @davidberenstein1957,

I was wondering if there were any changes in the past week. I had trained a model that was performing quite well with unseen data and when I loaded it this week it started acting up even with examples from the training data.

Thanks for your support.

[davidberenstein1957]

Hi,I reverted the config back into a solution that emperically proved more stable in my testing. I need to setup something more robust for this but will do so when I find the time. When I get back home, I will send you the old configs.CheersOn 20 Feb 2023, at 13:53, hectorhernandezbabbel @.***> wrote: Hi @davidberenstein1957, I was wondering if there were any changes in the past week. I had trained a model that was performing quite well with unseen data and when I loaded it this week it started acting up even with examples from the training data. Thanks for your support.

—Reply to this email directly, view it on GitHub, or unsubscribe.You are receiving this because you were mentioned.Message ID: @.***>

[hectorhernandezbabbel]

I'd really appreciate that. Thanks a lot!

[davidberenstein1957]

config = {
    "C": [1, 2, 5, 10, 20, 100],
    "kernel": ["linear", "rbf", "poly"],
    "max_cross_validation_folds": 5,
    "seed": None,
}

[hectorhernandezbabbel]

Thank you so much!

[hectorhernandezbabbel]

I have one more question @davidberenstein1957. Should this be included in the setup?

This is how I'm declaring it, but I don't think it's correct as I'm getting a ConfigValidationError´

nlp.add_pipe("text_categorizer", 
    config={
        "C": [1, 2, 5, 10, 20, 100],
        "kernel": ["linear", "rbf", "poly"],
        "max_cross_validation_folds": 5,
        "seed": None,
        "data": summary_tag,
        "model": "spacy",
         "multi_label": True,
    }
)

Thanks for the support again!