kaito
kaito copied to clipboard
Azure
Kubernetes AI Toolchain Operator
Please update the [Azure sku handler](https://github.com/kaito-project/kaito/blob/8f6f312d4f6fe7ededd33b40025dcb5019c58060/pkg/sku/azure_sku_handler.go) and proactively remove the skus impacted by [NV-series retirement](https://learn.microsoft.com/en-us/azure/virtual-machines/sizes/retirement/nv-series-retirement) and [NCv3, NC24rs retirement](https://learn.microsoft.com/en-us/azure/virtual-machines/ncv3-nc24rs-retirement) later this year.
**Is your feature request related to a problem? Please describe.** Presets and Kaito should have their own own release cadence. Benefits: - enable support for new models without having to...
Using the instructions in on this learn article fails to deploy the workspace. https://learn.microsoft.com/en-us/azure/aks/ai-toolchain-operator Create a new cluster using the instructions. Add --enable-oidc-issuer to the args then attempt to apply...
**Is your feature request related to a problem? Please describe.** Use AIKit to build customized model images from hf and use Kaito workspace to deploy the customized image for single...
quick start docs mention `Note that if you have installed another node provisioning controller that supports Karpenter-core APIs, the following steps for installing gpu-provisioner can be skipped.` it is unclear...
**Is your feature request related to a problem? Please describe.** **Describe the solution you'd like** inspired by https://github.com/NVIDIA/RULER. The effective context size is usually smaller than claimed length. It's a...
currently, tests require approval. since this is a CNCF project, this is a barrier for entry for new contributors and decrease pr merge velocity.
**Is your feature request related to a problem? Please describe.** Azure teams request for KAITO to validate and support node pool creation of Azure Linux 3.0 GPU nodes.
**Description:** To provide better default performance for distributed inference workloads, we should experiment with various parallelism strategies to identify a well-performing configuration that can be used as the out-of-the-box default....
**Is your feature request related to a problem? Please describe.** Fine tuning workspace should automatically get deleted once fine tuning is completed and kaito publishes adapter image in ACR. This...
