I was thinking to do 2. step manually

So you are saying that step 1 is the problematic one?

The latest code you've posted uplods to local.
I need to upload to my local Minio

So I need to fix Minio then?

artifact_object is local file in my case. I did not understand what you suggest

Hmm

The code took 0.16 for 33 MB

Oh, glad we found out a solution.

Does StorageManager.upload and upload_artifact use the same methods?

Is the only difference is task being async?

How about until weekend? Do you have a temporary solution=

Whose glitch on calculating the upload speed then? What will I do to fix my problem? 😞

When I give my Minio to output_uri argument, it uploads 500 KB /sec as before.

What if I register the artifact manually?

` from trains import Task

task = Task.init(project_name="SpeedTest",
task_name="test1",
task_type=Task.TaskTypes.data_processing,
output_uri="s3://<my_minio_ip>:<my_minio_port>/<my_bucket_name>",
reuse_last_task_id=False)

task.upload_artifact(name="test1",
artifact_object="<my_local_file>") `

Yep

Yes!
This solved the problem.
Thank you AgitatedDove14 !

Let's assume I've created a model with a task. When I execute a second task with same model name, is the model overridden or its version is automatically increased?

Thank you for your quick reply. AgitatedDove14
I've looked at the link you've provided. As I understood trains does not have auto versioning, it just infers the version from the task name right?

Okay, thanks!