Pierangelo Di Pilato

icon company Red Hat icon location Italy icon location Software Engineer | Knative Eventing TL | Cloud, Kubernetes & Kafka | Go, Scala & Java, Python

Results 462 comments of Pierangelo Di Pilato

Standardize discovery, installation and usage of LLM functions, agents and tools

@cali0707 presented a lighting talk at KubeCon featuring the prototype here: https://www.youtube.com/watch?v=wHxIaZawes8

Schduler: MAXFILLUP strategy will spread vreplicas across multiple pods

/cherry-pick release-1.15

Schduler: MAXFILLUP strategy will spread vreplicas across multiple pods

/cherry-pick release-1.16

Schduler: MAXFILLUP strategy will spread vreplicas across multiple pods

/cherry-pick release-1.14

[LLMInferenceServing] Set up envtest and suite for Controller Testing

@bartoszmajsak assigned to you as I know you have an early working setup/idea

[API] Define LLMInferenceService and LLMInferenceServiceConfig types and CRDs

/rerun-workflow "Python package"

[API] Define LLMInferenceService and LLMInferenceServiceConfig types and CRDs

/rerun-workflow E2T Tests

[API] Define LLMInferenceService and LLMInferenceServiceConfig types and CRDs

/rerun-workflow E2E Tests

[API] Define LLMInferenceService and LLMInferenceServiceConfig types and CRDs

/rerun-all

[API] Define LLMInferenceService and LLMInferenceServiceConfig types and CRDs

> Is distributed KV cache mandatory? I didn’t find a specific API to configure it, and based on the design doc, it seems to rely on the cache API. Just...