Gemini Tuning: TuningTask(start_time=None, complete_time=None)

Gemini API
, ,
1

I tried to follow the guide from Document to tune Gemini via python. I succifully added to the list of tuned models, however, after a few hours I tried typing

model = genai.get_tuned_model(f'tunedModels/generate-num-5283')
model

to check the status, it returns:

           source_model='models/gemini-1.0-pro-001',
           base_model='models/gemini-1.0-pro-001',
           display_name='',
           description='',
           temperature=0.9,
           top_p=1.0,
           top_k=0,
           state=<State.CREATING: 1>,
           create_time=datetime.datetime(2024, 5, 16, 6, 15, 29, 545843, tzinfo=datetime.timezone.utc),
           update_time=datetime.datetime(2024, 5, 16, 6, 15, 29, 545843, tzinfo=datetime.timezone.utc),
           tuning_task=TuningTask(start_time=None,
                                  complete_time=None,
                                  snapshots=[],
                                  hyperparameters=Hyperparameters(epoch_count=10,
                                                                  batch_size=4,
                                                                  learning_rate=1.0)))

Like it hasn’t started. Is this my problem? Or is it the problem of the API?

2

+1, I did the same, follow doc had the same issue. Left the model like that the next day it was trained, it took approximately 9 hours (time duration was written is aistudio). In studio it was “in pending” for hours.
This doesn’t seem to be normal behavior.

1 Like