Hello All. I Am Having Trouble Saving Pytorch Models As Clearml Artifacts When I Run My Training Script From Fir.Fire Calling.”

Answered

Hello all. I am having trouble saving PyTorch models as ClearML artifacts when I run my training script from fir.Fire calling.”
if __name__ == '__main__':
fire.Fire(run_my_trainer)
when I remove Fire scope all works correctly/

clearml 1.11.0
fire 0.5.0

  				
Posted 
	one year ago

					More
				  		
  Report
		
					EnormousGiraffe79
				
					0
					 × 1

Votes Newest

Answers 4

@<1523701087100473344:profile|SuccessfulKoala55>
No, I didn't see any errors. When I save model
torch.save(model.state_dict(), os.path.join(gettempdir(), " mny_model.pt "))
clearML task don't catch it when I run script from Fire

  				
Posted 
	one year ago

					More
				  		
  Report
		
					EnormousGiraffe79
				
					0
					 × 1

Sure Simple MNIST train toy example
@<1523701070390366208:profile|CostlyOstrich36>
if I use this run train without fire cli:

if __name__ == '__main__':
     main()

  				
Posted 
	one year ago

					More
				  		
  Report
		
					EnormousGiraffe79
				
					0
					 × 1

Hi @<1577830989026037760:profile|EnormousGiraffe79> , can you please a self contained code snippet that reproduces this?

  				
Posted 
	one year ago

					More
				  		
  Report
		
					CostlyOstrich36
				
					0

Hi @<1577830989026037760:profile|EnormousGiraffe79> , are you seeing any errors?

  				
Posted 
	one year ago

					More
				  		
  Report
		
					SuccessfulKoala55
				
					0
					 × 1

Write your answer

1K Views

4 Answers

one year ago