None

https://clear.ml/docs/latest/docs/configs/clearml_conf/

you can try to update the clearml.conf it has agent.cuda_version

you can use Task.update_output_model() to update the name of the output moel

https://clear.ml/docs/latest/docs/references/sdk/task#update_output_model

model*

it requires you set the weights and the framework name

ahh i see so one Task has multiple models that are trained

1000%

AgitatedDove14 yes it is a flask warning, i'm wondering why the clearml server isn't setup with wsgi?

if you go here https://clear.ml/docs/latest/docs/deploying_clearml/clearml_server_aws_ec2_ami you can see you can launch a community ami and if you go here https://clear.ml/docs/latest/docs/deploying_clearml/clearml_server_aws_ec2_ami/#clearml-server-aws-community-amis you will see the list of community AMI’s you can launch

i.e.
Task.update_output_model(name="custom_model")

from the clearml logs when setting up the clearml server using kubernetes

alternatively you could create a new OutputModel like here: https://github.com/allegroai/clearml/blob/master/examples/reporting/model_reporting.py , not sure if there is a way to stop the automatic uploading

@<1523701827080556544:profile|JuicyFox94> thanks for the reply, I’ll create a github issue 🙂 !

bugs i understand are better than ones i dont lol

Thanks!

import numpy as np
import matplotlib.pyplot as plt
import seaborn as sns
from clearml import Task

# Connecting ClearML with the current process,
# from here on everything is logged automatically
# Create a new task, disable automatic matplotlib connect
task = Task.init(project_name='Something', task_name='Something')

# Create plot and explicitly report as figure
N = 50
x = np.random.rand(N)
y = np.random.rand(N)
colors = np.random.rand(N)
area = (30 * np.random.rand(N))**2  # 0 to 15 ...

and when i’m running it in sagemaker

SuccessfulKoala55 do you mean request_handlers.py

not sure if it will work but its worth a try

SuccessfulKoala55 thanks alot for your help i greatly appreciate it, i think you are right that there is something happening with the request, I did some checking i can get it running on another computer (my non work computer), so i think you are right there must be something happening with my work related computer which is causing this to occur 😕 I'll keep trying to drill down on this issue on my work computer to see what the difference is, but I can atleast proceed now to as i g...

i think the call is wrong

sure where do i patch it?

` from argparse import ArgumentParser

from flask import Flask

from apiserver.config_repo import config
from apiserver.server_init.app_sequence import AppSequence
from apiserver.server_init.request_handlers import RequestHandlers

app = Flask(name, static_url_path="/static")
AppSequence(app).start(request_handlers=RequestHandlers())

=================== MAIN =======================

if name == "main":
p = ArgumentParser(description=doc)
p.add_argument(
"--po...

i would greatly appreciate that!

yeah

FLASK_DEBUG: "1" -> should 1 be a string or int?