allenai / tango

Organize your experiments into discrete steps that can be cached and reused throughout the lifetime of your research project.

Home Page:https://ai2-tango.readthedocs.io/

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Beaker Executor opens a ton of files

dirkgr opened this issue Β· comments

πŸ› Describe the bug

You have to do this in a catwalk context, on commit e7c5d158b9e8f1c925b4894037b5371a1efdeab7.

from tango import StepGraph
sg = StepGraph.from_file("experiments/everything/everything.jsonnet")
from tango import Workspace
ws = Workspace.from_url("beaker://ai2/task-complexity")
s = sg['trained_model_sciq_bert-base-uncased_1337']
r = ws.step_result(s)

At this point the Beaker Workspace will open millions of files, quickly exhausting the open file limit. Even when you Ctrl+C the proceedings, background threads keep trying to open files. Your console is now useless.

Versions

asdasdasd

I do this on a Mac. I think Linux has a higher default open file limit, so it takes a lot more to hit the same problem.

I realize (now) that step_result() is not the correct method to call. But the behavior is quite pathological.