worker: Add HF_TOKEN to runner container on creation

[hjpotter92]

Also provide clear hostname inside container when multiple containers are running on same host.

[codecov]

Codecov Report

❌ Patch coverage is 50.00000% with 3 lines in your changes missing coverage. Please review.
✅ Project coverage is 31.92539%. Comparing base (e7f0ee1) to head (7fc918b).
⚠️ Report is 1 commits behind head on master.

Files with missing lines	Patch %	Lines
ai/worker/docker.go	50.00000%	2 Missing and 1 partial ⚠️

Additional details and impacted files

@@                 Coverage Diff                 @@
##              master       #3693         +/-   ##
===================================================
+ Coverage   31.92386%   31.92539%   +0.00153%     
===================================================
  Files            156         156                 
  Lines          47441       47445          +4     
===================================================
+ Hits           15145       15147          +2     
- Misses         31401       31402          +1     
- Partials         895         896          +1     

Files with missing lines	Coverage Δ
ai/worker/docker.go	72.62774% <50.00000%> (-0.35020%)	⬇️

... and 4 files with indirect coverage changes

---

Continue to review full report in Codecov by Sentry.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update e7f0ee1...7fc918b. Read the comment docs.

Files with missing lines	Coverage Δ
ai/worker/docker.go	72.62774% <50.00000%> (-0.35020%)	⬇️

... and 4 files with indirect coverage changes

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[leszko]

Thanks for the PR @hjpotter92 Could you add in the description why we need to pass HF_TOKEN into the runner? Why does runner need to connect to HuggingFace?

[leszko]

Added one suggestion

[hjpotter92]

Thanks for the PR @hjpotter92 Could you add in the description why we need to pass HF_TOKEN into the runner? Why does runner need to connect to HuggingFace?

There's a possibility that this PR is not needed. @varshith15 mentioned getting rate limited. I'll add him as a reviewer as well and hold off on activity on this PR.

[varshith15]

so the main issue was that we are loading controlnets and unets in pytorch to detect the model type -- each model load takes about 3-4 sec and loading all controlnets adds upto the 15 sec josh noticed. this also adds to the vram as the pytorch models are not being unloaded after model type detection. ideally we shouldn't be loading any pytorch models and detect the model type from the model config instead (working on that).

so we should pause the PR for now, can revisit if we still face the issue after i push the above fix @hjpotter92