[5pt] Conduct experiments for TensorFlow 2.8 stacks

thoth-station / dependency-monkey-zoo

A set of configurations for Thoth's Dependency Monkey to resolve certain software stacks

GNU General Public License v3.0

0 stars 4 forks source link

[5pt] Conduct experiments for TensorFlow 2.8 stacks #14

Open fridex opened 2 years ago

fridex commented 2 years ago

Let's start testing TensorFlow 2.8 tests and let's see what issues we will observe:

run inspections for TensorFlow 2.8 stacks by triggering dependency monkey
check results of inspections and, eventually, derive prescriptions out of issues spotted

mayaCostantini commented 2 years ago

/assign @mayaCostantini /lifecycle active /sig stack-guidance

codificat commented 2 years ago

Adding triage labels: /triage accepted /priority important-soon

mayaCostantini commented 2 years ago

Acceptance criteria:

[ ] Prescriptions are generated for installation errors found during the experiment
[ ] If the number of errors found is consequent enough, write an article explaining how Dependency Monkey can help validate software in a cluster illustrated by this example

codificat commented 2 years ago

Example of similar previous work: resulted in https://thoth-station.ninja/j/tf_21_urllib3.html

In addition to the scripts we already have, TF documentation has some samples that could be reviewed and potentially incorporated to the things we test.

Another thing to do could be:

[ ] discuss findings with TF SIG build
[ ] scripts to test the application stack with are available at https://github.com/thoth-station/performance

sesheta commented 1 year ago

Issues go stale after 90d of inactivity. Mark the issue as fresh with /remove-lifecycle stale. Stale issues rot after an additional 30d of inactivity and eventually close.

If this issue is safe to close now please do so with /close.

/lifecycle stale

harshad16 commented 1 year ago

/remove-lifecycle stale /lifecycle frozen