Reputation
Badges 1
34 × Eureka!SuccessfulKoala55 we are looking into that as well .
The agent captures all output of a script, right?
We wrap everything in a tf.function()
like
` def f():
print('Tracing!')
tf.print('Executing')
tf.function(f)() `
I have to look deeper into our codebase to see what exactly happens.
It is the graph mode from TF1 and we use the mode still (in a session).
I will come up with a min working example (probably next week)
Does it make a difference for the autologging
if this is done in graphmode
?
Besides that, the procedure is the same, yes…
It makes training much faster
But the tasks should aways be connected to the same user, right?
Yes, we can freely create different users
Is there a way to join these accounts into a single one?
This would indicate the server created them on separate occasions
How can this be if the names are exactly the same?
What would happen if I deleted them out of the database?
SuccessfulKoala55 I updated the mongo db (removed 2 users) but I still see them in the User Filter in the WebUI. Does it take longer to update that or do I have to do that manually?
I navigated my way to the users in the database. The name
of User 1
is exactly the same, however, the instances have different IDs and credentials.
Thanks.
We’d like to consolidate these users in the WebUI because it’s very clustered. Is there a way to consolidate their experiments under the same user? I’d probably have to scroll through all experiments and change them manually, right?
Exactly. It’s hard to say for me whether this is an error on our end or yours. However, our setup is very rudimentary-->we basically use only the standard settings.
for to the database and check the
user
collection in the
auth
database
How do I find this?