Hi SourOx12
How do you set the iteration when you continue the experiment? is it with Task.init continue_last_task ?

AgitatedDove14
Yes, i use continue_last_task with reuse_last_task_id. The iteration number is the actual number of batches that were used, or the number of the epoch at which the training stopped. The iterations are served sequentially, but for some reason there is a gap in this picture

Hmm, I see the jump from 50 to 100, is that consistent with the last iteration on the aborted Task (before continuing )?

AgitatedDove14
The gap is always equal to the number of iterations completed before continuing training

SourOx12
Hmmm. So if last iteration was 75, the next iteration (after we continue) will be 150 ?

AgitatedDove14
Yes (if value of first iter is 0)

Okay let me check....

So the thing is clearml automatically detects the last iteration of the previous run, my assumption you also add it hence the double shift.
SourOx12 could that be it?

Sorry to answer so late AgitatedDove14
I also thought so and tried this thing:
` !pip install clearml
import clearml
id_last_start = '873add629cf44cd5ab2ef383c94b1c'

clearml.Task.set_credentials(...)
if id_last_start != '':

task = clearml.Task.get_task(task_id=id_last_start,project_name='tests', task_name='patience: 5 factor:0.5')

task = clearml.Task.init(project_name='Exp with ROP',
                         task_name='patience: 2 factor:0.75',
                         continue_last_task=True,
                         reuse_last_task_id=id_last_start,
                        )

else:
task = clearml.Task.init(project_name='tests', task_name='patience: 2 factor:0.75')
cfg.task = task
cfg.my_writer =task.get_logger()

def to_logger(path,step,val,cfg):
folder = path[:path.find('/')]
file = path[path.find('/')+1:]
if step == cfg.epoch:
step = step - cfg.start_epoch + 1

print(step,val)

    clearml.Logger.current_logger().report_scalar(folder, file, iteration=step, value=val)
elif step == cfg.step:
    step = step - cfg.start_step + 1 
    clearml.Logger.current_logger().report_scalar(folder, file, iteration=step, value=val) `And after restarting, I get these breaks in Scalars:  https://app.community.clear.ml/projects/2d68c58ff6f14403b51ff4c2d0b4f626/experiments/873add629cf44cd5ab2ef383c94b1c9b/output/execution

Hi SourOx12
I think that you do not actually need this one:
step = step - cfg.start_epoch + 1you can just do
step += 1ClearML Will take care of the offset itself

AgitatedDove14 This does not solve the problem unfortunately:( New exp: https://app.community.clear.ml/projects/2d68c58ff6f14403b51ff4c2d0b4f626/experiments/ec096e98ed5c4eccaf8047673023fc3e/output/execution
The image shows the eval log. The second column is val, the third column is step

I'm not sure I follow the example... Are you sure this experiment continued a previous run?
What was the last iteration on the previous run ?

AgitatedDove14
The last iteration before the restore was 2. Starting from the 3rd iteration, this is the restored model

AgitatedDove14
Can you please give some code examples where the training restore, because I haven't found any? I will be very grateful

Give me a minute, I'll check something

SourOx12
Run this example:
https://github.com/allegroai/clearml/blob/master/examples/reporting/scalar_reporting.py
Once, then change line #26 to:
task = Task.init(project_name="examples", task_name="scalar reporting", continue_last_task=True)and run again,

Hi AgitatedDove14 I finally found a solution to the problem. I should have written task.set_initial_iteration(0) after restore task. Thank you for your help

Nice SourOx12 !