Hi ! I Have Some Hyper Parameters Of My Clearml Task Which I Connected To Task With- Parameters = Task.Connect(Model_Train_Dict, Name='Train_Params'). I Ran The Task Manually From Vscode. Than, From Clearml Dashboard I Cloned It , Changed One Of The Para

Answered

Hi !
I have some hyper parameters of my clearml task which i connected to task with-
parameters = task.connect(model_train_dict, name='train_params').
I ran the task manually from vsCode. Than, from ClearML dashboard I cloned it , changed one of the parameters, and ran it with an agent (enqueue etc.) .
However, I don't see that the change done through dashboard is affecting the cloned task. It runs with the same parameter value as parent task.
Any idea why? Tnx

  				
Posted 
	one year ago

					More
				  		
  Report
		
					HollowPeacock58
				
					0
					 × 1

Votes Newest

Answers 2

Hi @<1571308003204796416:profile|HollowPeacock58> ! The changes should be reflected. Do you have a small example that could help us reproduce the issue?

  				
Posted 
	one year ago

					More
				  		
  Report
		
					SmugDolphin23
				
					0

looking now i think it's probably since on my code I first read parameters from config file-
config = HpsYaml(paras.config)
and after that, I set these parameters to dictionary, which I connect to the task-
hparams_dict = {'batch_size': config.hparas.batch_size,
'valid_step': config.hparas.valid_step,
'max_step': config.hparas.max_step}
parameters = task.connect(hparams_dict, name='hyper_params')
So maybe the parameters from the config file override the parameters setting from the GUI?
How is this suppose to be done in order for the GUI changed to apply?

  				
Posted 
	one year ago

					More
				  		
  Report
		
					HollowPeacock58
				
					0
					 × 1

Write your answer

996 Views

2 Answers

one year ago