
Reputation
Badges 1
383 × Eureka!Thanks CostlyOstrich36
And exact output after the package install and stuff:
Environment setup completed successfully Starting Task Execution: None
Thanks for the fast responses as usual AgitatedDove14 🙂
For different workloads, I need to habe different cluster scaler rules and account for different gpu needs
AS in clone the task from the UI and run it
As in if there are jobs, first level is new pods, second level is new nodes in the cluster.
It’s a task, it’s running in context of a project, but I don’t have a way to get the project name
Looks like Task.current_task() is indeed None in this case. Bit of log below where I print(Task.current_task()) as first step in the script
Environment setup completed successfully Starting Task Execution: None
But that itself is running in a task right?
How can a task running like this know its own project name?
Was asking about using iam roles without keys
Got the engine running.
curl <serving-engine-ip>:8000/v2/models/keras_mnist/versions/1
What’s the serving-engine-ip supposed to be?
I get other things from the project like the dataset
Ok couldn’t see it in the docs - https://clear.ml/docs/latest/docs/references/sdk/task
Lot of us are averse to using git repo directly
Oh yeah I wanted to update only for the gpu stuff for now
Yes, I have no experience with triton does it do lazy loading? Was wondering how it can handle 10s, 100s of models. If we load balance across a set of these engine containers with say 100 models and all of these models get traffic but distribution is not even, each of those engine container will load all those 100 models?
Yes, for datasets where we need GDPR compliance
Essentially - 1. run a task normally. 2. clone 3. edit to have only those two lines.
Question - since this is a task, why is Task.currnet_task() None?
AgitatedDove14 - i am disabling pytorch like above but still see auto models . I even see model import when running evaluation from a loaded model
The first line of it is getting a dataset and it’s failing with no project name
I just want to change git remote like https://gitserver.com/path/to.git -> mailto:git@gitserver.com :path/to.git
Great. Is there a good view of the roadmap?
One thing I am looking at is nbdev from fastai folks
Will try it out. Pretty impressed 🙂
Cool, didn't know it was disabled. This exact reason was why I created a wrapper over ClearML for my use so that people don't ever accidentally talk to demo server
Also is there a way to disable pytorch like this from clearml.conf?
I am going to be experimenting a bit as well, will get back on this topic in a couple of weeks 🙂