feat(gateway): route clef-flash to clef-serve

Browse Source

Restores the route removed when the model could not run. It runs now, on a different
runtime: SGLang needs v0.5.22+ which upstream publishes only on CUDA 13, and CUDA 13
dropped Volta kernels. clef-serve loads the same weights through plain transformers in
fp16, sharded over two V100s, and serves the model's own systemone() on /v1/systemone.

Weights measured at 17.75 GiB in BF16; fp16 costs the same bytes per parameter.

...

This commit is contained in:

Admin Bot

2026-10-29 23:06:32 -07:00

parent b9f469a92f

commit affb1caa63

1 changed files with 1 additions and 1 deletions

Show all changes Ignore whitespace when comparing lines Ignore changes in amount of whitespace Ignore changes in whitespace at EOL

Download Patch File Download Diff File Expand all files Collapse all files

k8s/gateway-config-secret.yaml

+1 -1

File diff suppressed because one or more lines are too long