llama32-3b deployment#11

Draft

juliaaano wants to merge 35 commits intomainfrom

Commits on Sep 11, 2025

testing subbing llama in for granite
taylorjordanNC
committed
stuff
taylorjordanNC
committed
fix taint
taylorjordanNC
committed
changing namespace issue
taylorjordanNC
committed
model name
taylorjordanNC
committed
commenting out max-model-len and max-number-seq
taylorjordanNC
committed
Revert "commenting out max-model-len and max-number-seq"
taylorjordanNC
committed
increasing --max-model-len and upping vllm version to 2.22 to match the granite servingruntime
taylorjordanNC
committed
Revert "increasing --max-model-len and upping vllm version to 2.22 to match the granite servingruntime"
taylorjordanNC
committed
increasing --max-model-len and upping vllm version to 2.22 to match the granite servingruntime
taylorjordanNC
committed
setting back / lowering the max-model-len and max-num-seq arguments to see if it was vllm version or expanded context window use
taylorjordanNC
committed
comment out max-num-seq and increase max-model-len
taylorjordanNC
committed
adjusting application3-scale.yaml so that it uses helm values instead of hardcoding the repoURL and targetRevision
taylorjordanNC
committed
changing activeDocOpenAPIRef to point to fork, adjusted cms-upload application boostrap to use helm values for repoURL and targetRevision
taylorjordanNC
committed
Changing max-num-seqs to 16
taylorjordanNC
committed
remove the llama guard deployment
taylorjordanNC
committed
changing git repo reference for the cms-upload values.yaml file
taylorjordanNC
committed
not hardcoding the repourl and target revision in part of 3scale
taylorjordanNC
committed

Commits on Sep 12, 2025

Commits on Sep 17, 2025

Commits on Sep 18, 2025

Use generic activedocs to fix validation issue
juliaaano
authored

Commits on Sep 24, 2025

Update inferenceservice.yaml - update to L40 GPU due to large model length
ritzshah
authored