I Want To Save Some Data On My Outputmodel In Order To Make It More Accessible When I'M Using The Model. When I Use

Answered

I want to save some data on my OutputModel in order to make it more accessible when I'm using the model. When I use myOuputModel.save_metadata("somekey", "someval", "str") I get failed validation when I try to upload the model.

jsonschema.exceptions.ValidationError: 'model' is a required property     

(...)
     'required': ['model', 'metadata'],
     'type': 'object'}

On instance:
    {'metadata': [{'key': 'final_score',
                   'type': 'float',
                   'value': '0.00)'}],
     'replace_metadata': False}

I've tried with both set_metadata and set_all_metadata(replace=False) but still get the same.

  				
Posted 
	one year ago

					More
				  		
  Report
		
					GrievingDuck15
				
					0
					 × 1

Votes Newest

Answers 4

Ok, the mistake was staring me in the eye the entire time.

I thought that set_metadata worked on the local output model. But what it actually does is try to update metadata on the remote. The output model doesn't exist on the remote before you call update_weights , so you have to set_metadata after upload.

  				
Posted 
	one year ago

					More
				  		
  Report
		
					GrievingDuck15
				
					0
					 × 1

I'll make one:)

  				
Posted 
	one year ago

					More
				  		
  Report
		
					GrievingDuck15
				
					0
					 × 1

import torch
import torchvision
from clearml import Task, OutputModel

scripted_model = torch.jit.script(torchvision.models.resnet18())
scripted_model.save("model.pt")

task = Task.init(project_name="OutputModelMVE", task_name="Testing shit")

outputmodel = OutputModel(task=task, framework='PyTorch')
outputmodel.update_labels({'Clearml': 0, 'Debug': 1})
outputmodel.set_upload_destination(INSERT DESTINATION)
metadatas = dict(Score=("0.2", "float"),
                SomeField=("SomeField", "str"),
                OtherField=("OtherField", "str"))

for key, (val, type) in metadatas.items():
    outputmodel.set_metadata(key, val, type, )

outputmodel.update_weights(weights_filename="model.pt",
                           target_filename="model.pt",
                           auto_delete_file=False,
                           is_package=True)
outputmodel.wait_for_uploads()

If you remove lines 17-18 and dont't change the metadtata the upload works.

  				
Posted 
	one year ago

					More
				  		
  Report
		
					GrievingDuck15
				
					0
					 × 1

Hi @<1562973083189383168:profile|GrievingDuck15> , do you have a standalone snippet that reproduces this error?

  				
Posted 
	one year ago

					More
				  		
  Report
		
					CostlyOstrich36
				
					0

Write your answer

1K Views

4 Answers

one year ago