011、AI大模型，GPU调度方法

字数统计: 117字   |   阅读时长≈ 1分

本人使用的显卡是： A100
使用AI大模型参数级别是： 100b

Ai大模型，GPU调度方法

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
kubectl label node gpu=a100


apiVersion: apps/v1
kind: StatefulSet
metadata:
  name: ollama
  namespace: llama
spec:
  serviceName: "ollama"
  replicas: 1
  selector:
    matchLabels:
      app: ollama
  template:
    metadata:
      labels:
        app: ollama
    spec:
      nodeSelector:
        gpu: a100
      containers:
        - name: ollama
          image: ollama/ollama:latest
          ports:
            - containerPort: 11434
          resources:
            requests:
              cpu: "6000m"
              memory: "6Gi"
              nvidia.com/gpu: "6"
            limits:
              cpu: "30000m"
              memory: "30Gi"
          volumeMounts:
            - name: ollama-volume
              mountPath: /root/.ollama
          tty: true
  volumeClaimTemplates:
    - metadata:
        name: ollama-volume
      spec:
        accessModes: ["ReadWriteOnce"]
        resources:
          requests:
            storage: 2000Gi