Aright, I might opt for just squashing the whole she-bang then.

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.

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",...

Where reregistering is uploading everything again?

I'll make one:)

So I tested using the version that I can see in the request going to the server, but the api returns "Requested version: 2.28, Used version 1.0" for some reason. I'm not sure if the message is wrong or it actually used 1.0.