and then have a wrapper that gets the model data and selects which way to construct and deserialise the model class.
` def get_model(task_id, model_name):
task = Task.get_task(task_id)
try:
model_data = next(model for model in task.models['output'] if model.name == model_name)
except StopIteration as ex:
raise ValueError(f'Model {model_name} not found in: {[model.name for model in task.models["output"]]}')
filename = model_data.get_local_copy()
model_type =...

while in our own code:
if model_type == 'XGBClassifier': model = XGBClassifier() model.load_model(filename)

Yes, this is exactly how I solved it at the end

pickle.dump({ 'model': model, 'X_train': X_train, 'Y_train': Y_train, 'X_test': X_test, 'Y_test': Y_test, 'impute_values': impute_values }, open(self.output_filename, 'wb'))

That's because then I need to teach every DS how to use the ClearML api

this is a bit WIP but we save it with the design of the model:
` parameters = dict(self.parameters, model_type='XGBClassifier')
...

output_model.update_design(config_dict=parameters) `

Is there any restriction on what can be in preprocess.py?

import xgboost # noqa self._model = xgboost.Booster() self._model.load_model(self._get_local_model_file())

actually it is this:
https://github.com/allegroai/clearml-serving/blob/219fa308df2b12732d6fe2c73eea31b72171b342/clearml_serving/serving/preprocess_service.py#L435

I think this is because of the version of xgboost that serving installs. How can I control these?

Is there a more detailed logging about what's going on?

This is very different than from in the xgboost example:

Python 3.8.13

apparently it did not work

I want the model to be stored in a way that clearml-serving can recognise it as a model

So the only thing left is models

wait

The DSes would expect the same interface as they used in the code that saved the model (me too TBH)

I'd rather just fail if they try to use a model that is unknown.

I am not sure I want to do that

I just disabled all of them with auto_connect_frameworks=False

I didn't realise that pickling is what triggers clearml to pick it up. I am actually saving a dictionary that contains the model as a value (+ training datasets)

well, it just shoved the dataset files with cryptic data_1/data_2 names among the artifacts.

This receives the payload from the server and turns it into something that can be fed to the model. This in out case depends on a data structure that is stored on the clearml server as an artifact. I need to communicate this to the class so it can pick it up and use it when called

yeah, Friday afternoon

BTW you are not exporting Framework in __ init __ so you need to import it like from clearml.model import Framework

If I do this it still autorecords the sklearn one

I am running a script

also random tasks are popping up in the DevOps project in the UI

I absolutely need to pin the packages (incl main DS packages) I use.

Having human readable ids always help communication but programmatically we definitely going to use the "real" id. But I think we are too early into this and I will report back on how it is going with this.