Label Studio AssertionError while Submitting Annotation [label_studio_ml.model::get_result::58]

Hi, I have implemented a text classification task with two class classifications and connected label-studio-ml-backend with the label-studio. The Training and Prediction Retrieval task is being done correctly. The only issue is is retraining after each annotation submission. It is throwing the following error

[2022-08-21 01:20:41,152] [ERROR] [label_studio_ml.model::get_result::58] 
Traceback (most recent call last):
  File "...\anaconda3\envs\ls\lib\site-packages\label_studio_ml\model.py", line 56, in get_result
    job_result = self.get_result_from_job_id(model_version)
  File "...\anaconda3\envs\ls\lib\site-packages\label_studio_ml\model.py", line 110, in get_result_from_job_id
    assert isinstance(result, dict)

ML Model Setting is as follows

Seems like get_result method in JobManager Class is not able to get the annotation result.

This is also sending empty data in fit(self, completions, workdir=None, **kwargs) method as shown below. I have used print statements to print these on the console for debugging.

***************-------FIT -----***********
completions -> ()
***************-------END-----***********

Label set has been changed:[]

***************-------FIT -----***********
input_text []
output_labels []
output_labels_idx []
***************-------END-----***********

Since input data is empty therefore fit is throwing the following error while doing model.fit()

ValueError: empty vocabulary; perhaps the documents only contain stop words

Any word of advice would be appreciable!

[KonstantinKorotaev]

Hi @askgitqa When you are using active learning your fit method get only an event that you shoud handle. Check this example.

[KonstantinKorotaev]

This issue is still unresolved [label_studio_ml.model::get_result::58]

This issue is not connected with train issue, it tells that some of your training sessions ended without result and can’t be used as checkpoint.