1. LLM Deployment Tool llama.cpp Research on large models is split into two parts: training and inference. The training process is essentially the process of finding model parameters that minimize the model’s loss function and optimize the inference results. Once training is complete, the model’s parameters are fixed, and at

1. What Problem Argo CD Solves 1.1 Starting from GitOps GitOps originated from a blog post published by Weaveworks in 2017: GitOps - Operations by Pull Request. In it, Alexis introduced a deployment approach that treats Git as the single source of truth. In GitOps practice, we need to define software infrastructure in a Git repository for management.

A Deployment controls the number and state of Pods by creating ReplicaSets. This post mainly introduces some common Deployment operations. 1. Deployment yaml Format Passing the --dry-run flag means the command is not actually executed, only YAML output is generated: 1 kubectl create deployment nginx --image=nginx --dry-run -o yaml apiVersion: apps/v1 kind: Deployment metadata: creationTimestamp: null labels: app: nginx name: nginx spec: replicas: 1 selector: matchLabels: app: nginx strategy: {} template: metadata: creationTimestamp: null labels: app: nginx spec: containers: - image: nginx name: nginx resources: {} status: {} 2.

1. Why Helm Is Needed One important design philosophy in Kubernetes is declarative operation. Users change the system by setting the system’s expected state. For example, the current replica count is 2 and it needs to be adjusted to 3. The declarative way is to modify the replica count in

I have been learning Go recently, and the internal PaaS platform we commonly use happens to support Go as well as related Web frameworks. A PaaS system that supports multiple languages cannot do without the buildpack mechanism. Although PaaS platforms keep being upgraded, the buildpack mechanism has always been retained.