From 8404b070f3d61c806fa6b35912a830bc24dddcbe Mon Sep 17 00:00:00 2001 From: Volodymyr Tsap Date: Thu, 16 Nov 2023 10:52:50 +0000 Subject: [PATCH] Documentation init --- README.md | 48 +++++++++++++++++++++++++++++++++++ huggingface-model/Chart.yaml | 2 +- huggingface-model/README.md | 9 +++++-- huggingface-model/values.yaml | 12 ++++----- 4 files changed, 62 insertions(+), 9 deletions(-) create mode 100644 README.md diff --git a/README.md b/README.md new file mode 100644 index 0000000..707de52 --- /dev/null +++ b/README.md @@ -0,0 +1,48 @@ +# The SHALB Helm Library + +Here we collect useful charts we designed in [SHALB](https://shalb.com) also used in our [cluster.dev](htts://cluster.dev) product for building cloud installers and creating infrastructure templates with Terraform modules and Helm charts. + +## Before you begin + +### Prerequisites + +- Kubernetes 1.23+ +- Helm 3.8.0+ + +## Setup a Kubernetes Cluster + +The quickest way to setup a Kubernetes cluster in different clouds to install SHALB Charts is by using [cluster.dev](https://docs.cluster.dev). + +## Bootstrapping Kubernetes in Different Clouds + +Create fully featured Kubernetes clusters with required addons: + +| Cloud Provider | Kubernetes Type | Sample Link | Technologies | +|----------------|-----------------|-------------------------|------------------| +| AWS | EKS | [**AWS-EKS**](https://docs.cluster.dev/examples-aws-eks/) | | +| AWS | K3s | [**AWS-K3s**](https://docs.cluster.dev/examples-aws-k3s/) | | +| GCP | GKE | [**GCP-GKE**](https://docs.cluster.dev/examples-gcp-gke/) | | +| AWS | K3s + Prometheus| [**AWS-K3s Prometheus**](https://docs.cluster.dev/examples-aws-k3s-prometheus/) | | +| DO | K8s | [**DO-K8s**](https://docs.cluster.dev/examples-do-k8s/) | | + +## Using Helm + +To install Helm, refer to the [Helm install guide](https://github.com/helm/helm#install) and ensure that the helm binary is in the PATH of your shell. +Once you have installed the Helm client, you can deploy a SHALB Helm Chart into a Kubernetes cluster. +Please refer to the [Quick Start guide](https://helm.sh/docs/intro/quickstart/). + +## Using Helm with cluster.dev + +Example of how to deploy application with Helm and Terraform to Kubernetes: + +| Description | Sample Link | Technologies | +|-----------------------------|---------------------------------------|------------------| +| Kubernetes Terraform Helm | [**Quick Start with Kubernetes**](https://docs.cluster.dev/get-started-cdev-helm/) | | + +## License + +Copyright © 2023 SHALB. + +Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with the License. You may obtain a copy of the License at + +http://www.apache.org/licenses/LICENSE-2.0 \ No newline at end of file diff --git a/huggingface-model/Chart.yaml b/huggingface-model/Chart.yaml index c6c740a..ceda344 100644 --- a/huggingface-model/Chart.yaml +++ b/huggingface-model/Chart.yaml @@ -1,6 +1,6 @@ apiVersion: v2 name: huggingface-model -description: Helm chart for deploy Hugging Face to kubernetes cluster. See [Hugging Face models](https://huggingface.co/models) +description: Helm chart for deploy Hugging Face models and chat-ui to Kubernetes cluster. See [Hugging Face models](https://huggingface.co/models) type: application diff --git a/huggingface-model/README.md b/huggingface-model/README.md index 85aae8a..20d4f73 100644 --- a/huggingface-model/README.md +++ b/huggingface-model/README.md @@ -1,6 +1,11 @@ -# Helm chart for deploy Hugging Face to kubernetes cluster +# Helm chart for deploy HuggingFace models to Kubernetes cluster -See [Hugging Face models](https://huggingface.co/models) +Charts install [Text Generation Inference](https://github.com/huggingface/text-generation-inference) container and serves [Text Generation LLM models](https://huggingface.co/models?pipeline_tag=text-generation). +It is possible to inject another image to serve inference with different approach. + +init-container is used to download model to PVC storage from HuggingFace directly or from s3-compatible(and from other storage). + +Also it would deploy [HuggingFace chat-ui](https://github.com/huggingface/chat-ui) image and configure it to use with deployed model to be able to chat with it in browser. ## Parameters diff --git a/huggingface-model/values.yaml b/huggingface-model/values.yaml index 6efd659..907c6f2 100644 --- a/huggingface-model/values.yaml +++ b/huggingface-model/values.yaml @@ -3,15 +3,15 @@ ## ref: https://huggingface.co/models ## @param model.organization Models' company name on huggingface, required! ## @param model.name Models' name on huggingface, required! -## e.g. to deploy model https://huggingface.co/segmind/SSD-1B use configuration below: -## organization: segmind -## name: SSD-1B +## e.g. to deploy model https://huggingface.co/HuggingFaceH4/zephyr-7b-beta use configuration below: +## organization: HuggingFaceH4 +## name: zephyr-7b-beta ## model: organization: "" name: "" -## Init configuration. By default, init clone model from huggingface git. +## Init configuration. By default, init clone model from Huggingface git using git-lfs. ## The another way is to upload model to s3 bucket to reduce init delay and external traffic. ## @param init.s3.enabled Turn on/off s3 data source Default: disabled ## @param init.s3.bucketURL Full s3 URL included path to model's folder @@ -19,9 +19,9 @@ model: init: s3: enabled: false - bucketURL: s3://k8s-model-zephyr/llm/deployment/segmind/SSD-1B + bucketURL: s3://k8s-model-zephyr/llm/deployment/HuggingFaceH4/zephyr-7b-beta -## huggingface block configure running text-generation-launcher internal port and additional arguments +## Huggingface block configure running text-generation-launcher internal port and additional arguments ## @param huggingface.containerPort Deployment/StatefulSet ContainerPort, optional ## huggingface: