[BUG] "GET /is_model_training HTTP/1.1" 200 OK stuck about autotrain-advanced

[abhishekkrthakur]

is_model_training will continuously check if a model is training. its not an issue. does your training fail? do you have any other errors/issues in the logs?

[abhishekkrthakur]

then its done

[leovfxuk]

How long should this loop run for? I'm getting the same thing for hours. There's no percentage read out. Last night I had to leave it running for so long that I went to bed and came back and colab session had ended. Don't know if the loop ever finished or not.

[abhishekkrthakur]

is_model_training is polled continuously. if your model is trained, you find it on hf hub or locally. the logs/errors are also displayed in the same output where you see is_model_training

[leovfxuk]

OK Thanks. So if I have a file: pytorch_lora_weights.safetensors then the training is finished and I can end the runtime even though it's still displaying "GET /is_model_training HTTP/1.1" 200 OK? Or should I wait until it says the model isn't training?

[abhishekkrthakur]

yes!

[abhishekkrthakur]

does the UI say model is training?

[leovfxuk]

yes

"GET /is_model_training HTTP/1.1" 200 OK

[leovfxuk]

I should wait until this finishes?

Am i just being impatient?

[abhishekkrthakur]

i meant in the ui. what does it say here?

[leovfxuk]

OK. That says "no running jobs"

[leovfxuk]

thanks