Skip to content

kev-deploy and kev-finetune: pin KEV_REF to fc4a17e (device probe in a child process #218 #231, compare null diagnostics #217) #233

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
jaredpalmer merged 1 commit into from
Oct 10, 2026
Merged

kev-deploy and kev-finetune: pin KEV_REF to fc4a17e (device probe in a child process #218 #231, compare null diagnostics #217) #233

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion skills/kev-deploy/scripts/kev_serve.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -28,7 +28,7 @@
import modal
import modal.experimental

KEV_REF = "84847f0a883d900f7de5b7a57eaa341ca7f9a6b4" # github.com/jaredpalmer/kev commit whose kev package this endpoint runs
KEV_REF = "fc4a17e194bca35a48c57211392ec49ea54a94ac" # github.com/jaredpalmer/kev commit whose kev package this endpoint runs
# GPU preference lists (Modal takes the first with capacity), from runs/serve-*/, runs/fused-27b-*/ and runs/serving-*/report.json in the repo.
# An L4 is enough for the 0.8B but runs out of compute on the 4B; the L40S is the cheapest GPU that answers the 4B in tens
# of milliseconds, the H100 the fastest for the 4B and 9B. The A100 is slower than the L40S here and costs more. Kev-27B
Expand Down
2 changes: 1 addition & 1 deletion skills/kev-finetune/scripts/kev_modal.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -29,7 +29,7 @@
# Launch-time settings that the container must see identically: they travel in the image env (names only, never secret
# values). The module is re-evaluated inside the container, and a Secret list or a served run that differs there either
# fails the container ("Function has N dependencies but got M") or serves the wrong model.
SETTINGS = {"KEV_APP_NAME": "kev-finetune", "KEV_REF": "84847f0a883d900f7de5b7a57eaa341ca7f9a6b4", "KEV_SERVE_RUN": "jaredpalmer/kev-4b", "KEV_HF_SECRET": "", "KEV_SERVE_SECRET": ""}
SETTINGS = {"KEV_APP_NAME": "kev-finetune", "KEV_REF": "fc4a17e194bca35a48c57211392ec49ea54a94ac", "KEV_SERVE_RUN": "jaredpalmer/kev-4b", "KEV_HF_SECRET": "", "KEV_SERVE_SECRET": ""}
SETTINGS = {k: os.environ.get(k, v) for k, v in SETTINGS.items()}
APP_NAME, KEV_REF, SERVE_RUN = SETTINGS["KEV_APP_NAME"], SETTINGS["KEV_REF"], SETTINGS["KEV_SERVE_RUN"]
KEV_REPO = "https://github.com/jaredpalmer/kev.git"
Expand Down
Loading