new

gpuaas/jupyter_kustomize/base/tf.yaml

@@ -32,7 +32,7 @@ spec:
           image: tensorflow/tensorflow:2.16.2-gpu-jupyter
           resources:
             limits:
-              nvidia.com/gpu: 1
+              nvidia.com/gpu: 2
           ports:
             - containerPort: 8888
               name: notebook

gpuaas/jupyter_kustomize/overlays/7_jupyter/kustomization.yaml

@@ -0,0 +1,16 @@
+apiVersion: kustomize.config.k8s.io/v1beta1
+kind: Kustomization
+
+resources:
+  - ../../base
+
+namePrefix: a7-
+namespace: org1
+  
+
+patches:
+  - path: patch-pvc.yaml
+    target:
+      kind: PersistentVolumeClaim
+      name: tf-notebook-pvc
+      namespace: org1

gpuaas/jupyter_kustomize/overlays/7_jupyter/patch-pvc.yaml

@@ -0,0 +1,10 @@
+apiVersion: v1
+kind: PersistentVolumeClaim
+metadata:
+  name: tf-notebook-pvc
+  namespace: org1
+spec:
+  resources:
+    requests:
+      storage: 10Gi
+

gpuaas/jupyter_kustomize/overlays/7_jupyter/test.yaml

@@ -0,0 +1,46 @@
+apiVersion: v1
+kind: PersistentVolumeClaim
+metadata:
+  name: a1-tf-notebook-pvc
+  namespace: org2
+spec:
+  accessModes:
+  - ReadWriteOnce
+  resources:
+    requests:
+      storage: 10Gi
+  storageClassName: nfs-client
+---
+apiVersion: apps/v1
+kind: Deployment
+metadata:
+  name: a1-tf-notebook
+  namespace: org2
+spec:
+  replicas: 1
+  selector:
+    matchLabels:
+      app: tf-notebook
+  template:
+    metadata:
+      labels:
+        app: tf-notebook
+    spec:
+      containers:
+      - image: tensorflow/tensorflow:2.16.2-gpu-jupyter
+        name: tf-notebook
+        ports:
+        - containerPort: 8888
+          name: notebook
+        resources:
+          limits:
+            nvidia.com/gpu: 1
+        volumeMounts:
+        - mountPath: /sample
+          name: notebook-storage
+      nodeSelector:
+        nodegroup: gpu
+      volumes:
+      - name: notebook-storage
+        persistentVolumeClaim:
+          claimName: a1-tf-notebook-pvc

monitoring/scp.sh

@@ -2,8 +2,8 @@
 
 # 변수 정의
 NAMESPACE="org1"       # 예: org1
-POD_NAME="a2-tf-notebook-6bd88b8884-582nr"         # 예: tf-notebook-abcdef
-LOCAL_FILE="./sample.py"   # 예: ./config.json
+POD_NAME="a7-tf-notebook-5b5ddb5f7-mmlqv"         # 예: tf-notebook-abcdef
+LOCAL_FILE="./unlimit.py"   # 예: ./config.json
 TARGET_PATH="/tf"       # 예: /app/config/
 
 # 멀티컨테이너가 아닐 경우

monitoring/unlimit.py

@@ -0,0 +1,47 @@
+import tensorflow as tf
+from tensorflow.keras import layers, models
+import numpy as np
+import time
+
+# 모델 평가 시작 시간 기록
+start_time = time.time()
+
+# 데이터셋 로드 및 전처리
+(x_train, y_train), (x_test, y_test) = tf.keras.datasets.mnist.load_data()
+
+# 데이터 정규화 및 차원 확장
+x_train = x_train / 255.0
+x_test = x_test / 255.0
+x_train = np.expand_dims(x_train, -1)
+x_test = np.expand_dims(x_test, -1)
+
+# 간단한 CNN 모델 정의
+model = models.Sequential([
+    layers.Conv2D(32, (3, 3), activation='relu', input_shape=(28, 28, 1)),
+    layers.MaxPooling2D((2, 2)),
+    layers.Conv2D(64, (3, 3), activation='relu'),
+    layers.MaxPooling2D((2, 2)),
+    layers.Flatten(),
+    layers.Dense(64, activation='relu'),
+    layers.Dense(10, activation='softmax')
+])
+
+# 모델 컴파일
+model.compile(optimizer='adam',
+              loss='sparse_categorical_crossentropy',
+              metrics=['accuracy'])
+
+# 모델 요약
+model.summary()
+
+# 모델 훈련
+model.fit(x_train, y_train, epochs=500, batch_size=256, validation_split=0.2)
+
+# 모델 평가
+test_loss, test_acc = model.evaluate(x_test, y_test, verbose=2)
+print(f"Test accuracy: {test_acc:.2f}")
+
+# 모델 평가 종료 시간 기록 및 출력
+end_time = time.time()
+print(f"Evaluation time: {end_time - start_time:.2f} seconds")
+