oeDeploy Use Cases

Use Case 1: One-Click Kubernetes Deployment

Prepare three virtual machines with 2 cores and 4 GB memory (with Layer-3 network connectivity), using any version of openEuler 24.03 or 22.03. The goal is to deploy a Kubernetes cluster consisting of 1 master and 2 worker nodes.
On any node, install the oeDeploy CLI tool oedp (default integrated in openEuler software sources released in 2025 and later).
bash
```
yum install -y oedp
```
Execute the following commands to obtain the plugin package, ensuring the directory kubernetes-1.31.1 appears in the current directory.
shell
```
oedp repo update
oedp init kubernetes-1.31.1
```
Execute the info command to view plugin details.
shell
```
oedp info -p kubernetes-1.31.1
```

Modify the project configuration file with actual node information.

shell

vim kubernetes-1.31.1/config.yaml

yaml

all:
  children:
    masters:
      hosts:
        # The number of master nodes must be odd
        172.27.76.114:                     # master node IP
          ansible_host: 172.27.76.114      # master node IP
          ansible_port: 22
          ansible_user: root
          ansible_password: ""
          architecture: amd64              # [ amd64, arm64 ]
          oeversion: 24.03-LTS             # [ 22.03-LTS, 24.03-LTS ]
          runtime: docker                  # [ docker, containerd ]
    workers:
      hosts:
        172.27.70.60:                      # worker node IP
          ansible_host: 172.27.70.60       # worker node IP
          ansible_port: 22
          ansible_user: root
          ansible_password: ""
          architecture: amd64
          oeversion: 24.03-LTS
          runtime: docker
        172.27.72.90:
          ansible_host: 172.27.72.90
          ansible_port: 22
          ansible_user: root
          ansible_password: ""
          architecture: amd64
          oeversion: 24.03-LTS
          runtime: docker
    new-workers:  # The new-workers field cannot be deleted
      hosts:
  vars:
    ansible_ssh_common_args: '-o StrictHostKeyChecking=no'

    init_cluster_force: "true"             # e.g. [ "true", "false" ]  Force cluster initialization
    remove_master_no_schedule_taints: "true"
    service_cidr: 10.96.0.0/16             # Service CIDR
    pod_cidr: 10.244.0.0/16                # Pod IP CIDR
    certs_expired: 3650                    # Certificate expiration days
    # lb_kube_apiserver_ip:
    lb_kube_apiserver_port: 8443
    has_deployed_containerd: "false"       # e.g. [ "true", "false" ]  Whether containerd is already deployed

    # The following parameters must match those in host.ini during the build process; ignore if not involved
    kubernetes_version: 1.31.1
    calico_version: 3.28.2
    pause_image: "registry.k8s.io/pause:3.10"

Note: Ensure SSH connectivity between nodes, supporting both password and key-based authentication. If using key-based authentication, password configuration is not required.

Execute the following command to start automated deployment.
shell
```
oedp run install -p kubernetes-1.31.1
```
Execute the following command to uninstall Kubernetes.
shell
```
oedp run uninstall -p kubernetes-1.31.1
```
The -p parameter indicates the extracted plugin directory. If you enter the kubernetes-1.31.1 plugin root directory, the -p parameter is not required when executing oedp commands.

Use Case 2: One-Click AnythingLLM Deployment

Install the oeDeploy CLI tool oedp (default integrated in openEuler software sources released in 2025 and later).
bash
```
yum install -y oedp
```
Obtain and initialize the AnythingLLM plugin.
shell
```
oedp repo update
oedp init anythingLLM
```
Modify config.yaml according to your actual situation. Ensure the deployment node has normal network connectivity.

One-click deployment:

bash

oedp run install -p anythingLLM  # -p <plugin directory>

One-click uninstallation:

bash

oedp run uninstall -p anythingLLM  # -p <plugin directory>

Demo: Deploy and Print AnythingLLM Information

Edit the config.yaml file.

yaml

all:
hosts:
  host1:
    ansible_host: 192.168.124.17   # Remote IP
    ansible_port: 22               # Port number
    ansible_user: root             # Username
    ansible_password: ""       # Password

vars:
  docker_download_url:
    x86_64: https://download.docker.com/linux/static/stable/x86_64/docker-26.1.4.tgz
    aarch64: https://download.docker.com/linux/static/stable/aarch64/docker-26.1.4.tgz
  docker_download_path: /tmp
  docker_image_urls: https://docker.m.daocloud.io

  container_name: "anythinglLLM-container"
  docker_image: "mintplexlabs/anythingllm"
  storage_location: "{{ lookup('env', 'HOME') }}/anythingllm"
  host_port: 3001

Enter the container and test the AnythingLLM service.
- Enter the container:
bash
```
docker exec -it anythinglLLM-container /bin/bash
```
- Test the AnythingLLM service:
bash
```
curl localhost:3001
```
Test the AnythingLLM service in the deployment environment.
bash
```
-curl 192.168.124.17:3001/ curl localhost:3001
```

Normal response is as follows:

html

<!DOCTYPE html>
<html lang="en">
  <head>
    <meta charset="UTF-8" />
    <meta name="viewport" content="width=device-width, initial-scale=1.0" />
    <link type="image/svg+xml" href="/favicon.png" >
    <title >AnythingLLM | Your personal LLM trained on anything</title>
    <meta name="title" content="AnythingLLM | Your personal LLM trained on anything" >
    <meta description="title" content="AnythingLLM | Your personal LLM trained on anything" >
    <meta property="og:type" content="website" >
    <meta property="og:url" content="https://anythingllm.com" >
    <meta property="og:title" content="AnythingLLM | Your personal LLM trained on anything" >
    <meta property="og:description" content="AnythingLLM | Your personal LLM trained on anything" >
    <meta property="og:image" content="https://raw.githubusercontent.com/Mintplex-Labs/anything-llm/master/images/promo.png" >
    <meta property="twitter:card" content="summary_large_image" >
    <meta property="twitter:url" content="https://anythingllm.com" >
    <meta property="twitter:title" content="AnythingLLM | Your personal LLM trained on anything" >
    <meta property="twitter:description" content="AnythingLLM | Your personal LLM trained on anything" >
    <meta property="twitter:image" content="https://raw.githubusercontent.com/Mintplex-Labs/anything-llm/master/images/promo.png" >
    <link rel="icon" href="/favicon.png" >
    <link rel="apple-touch-icon" href="/favicon.png" >
    <script type="module" crossorigin src="/index.js"></script>
    <link rel="stylesheet" href="/index.css">
  </head>
  <body>
    <div id="root" class="h-screen"></div>
  </body>
</html>

Use Case 3: One-Click DeepSeek-R1 Deployment

Prepare an openEuler environment (22.03-LTS-SPX, 24.03-LTS-SPX). The recommended specification for the 8B model is greater than 8U16G.
Install the oeDeploy CLI tool oedp (default integrated in openEuler software sources released in 2025 and later).
bash
```
yum install -y oedp
```
Obtain and initialize the DeepSeek-R1 plugin.
bash
```
oedp repo update
oedp init deepseek-r1
```

Modify deepseek-r1/config.yaml according to your actual situation. Users can directly use the example here without any modifications.

yaml

all:
  hosts:
    # Choose either local deployment or remote deployment
    # ================ Local Deployment =====================
    localhost:
      ansible_connection: local
    # ================ Remote Deployment =====================
    # host1:
    #   ansible_host: 127.0.0.1      # Remote IP
    #   ansible_port: 22             # Port number
    #   ansible_user: root           # Username
    #   ansible_password: ""   # Password
  vars:
    deepseek_version: 8b
    # Official ollama download URL: https://ollama.com/download/ollama-linux-amd64.tgz, note the distinction between amd64 and arm64
    # To improve download speed, it has been temporarily stored on the OEPKGS server
    ollama_download: https://repo.oepkgs.net/openEuler/rpm/openEuler-24.03-LTS/contrib/oedp/2025.0330/ollama-linux-amd64.tgz
    ollama_download_path: /tmp     # Target download path
    # Model file download URL: https://www.modelscope.cn/models/unsloth/DeepSeek-R1-Distill-Llama-8B-GGUF/resolve/master/DeepSeek-R1-Distill-Llama-8B-Q4_K_M.gguf
    # To improve download speed, it has been temporarily stored on the OEPKGS server
    modelfile_download: https://repo.oepkgs.net/openEuler/rpm/openEuler-24.03-LTS/contrib/oedp/2025.0330/DeepSeek-R1-Distill-Llama-8B-Q4_K_M.gguf
    modelfile_download_path: /tmp  # Target download path
    # Model parameters
    parameter:
      temperature: 0.7
      top_p: 0.7
      top_k: 30
      num_ctx: 4096
      num_thread: 8    # Number of threads, recommended not to exceed CPU cores
      num_gpu: 0       # Number of GPUs: 0 for none, -1 for all.
    
    ansible_ssh_common_args: '-o StrictHostKeyChecking=no'

Trigger one-click automated DeepSeek-R1 deployment. The oeDeploy tool will download the required files, automatically install and deploy, and complete the configuration for DeepSeek-R1.
bash
```
oedp run install -p deepseek-r1  # -p <path after extracting tar.gz>
```
On the node after deployment is complete, open the DeepSeek-R1 interactive terminal to start a conversation.
bash
```
ollama run deepseek-r1:8b
```

Use Case 4: One-Click KubeRay Deployment

Environment Preparation

Deploy a Kubernetes cluster:
Deploy Kubernetes 1.31.1 on the target cluster machines. Here we take a standard K8S cluster with 1 master + 2 workers as an example, where the master node IP is 192.168.1.101. You can use the kubernetes-1.31.1 plugin for automated deployment.
Install Helm:
Install Helm on the master node. You can use the helm-3.9.0 plugin for automated installation.

Deploy KubeRay

Obtain the kuberay plugin:
Use the following commands to obtain and initialize the kuberay plugin.
shell
```
oedp repo update
oedp init kuberay
```

Modify config.yaml:

The following is an example of kuberay/config.yaml. Modify this file according to your actual situation.

yaml

all:
  hosts:
    host1:
      ansible_host: 192.168.1.101      # Master node IP
      ansible_port: 22                 # SSH port
      ansible_user: root               # SSH user, non-root users require sudo privileges
      ansible_password: ""       # Password of the above user
  vars:
    temp_path: /tmp
    namespace: kuberay                 # Namespace
    version: 1.2.2                     # KubeRay chart version

    helm_repo_name: "kuberay"          # Local Helm repository alias
    helm_repo_url: "https://repo.oepkgs.net/openEuler/rpm/openEuler-24.03-LTS/contrib/oedp/2025.0330/kuberay-helm"

    kuberay_operator_values_file: ""
    # The above configuration item kuberay_operator_values_file is an advanced configuration: specify a values file for customizing
    # the installation of the KubeRay operator component. Provide an absolute path; if left empty, the following configuration will be used:
    kuberay_operator_values:
      repository: "hub.oepkgs.net/oedeploy/quay.io/kuberay/operator" # KubeRay operator docker image URL
      tag: v1.2.2                      # KubeRay operator docker repository tag

    ray_cluster_values_file: ""
    # The above configuration item ray_cluster_values_file is an advanced configuration: specify a values file for customizing
    # the installation of the Ray component. Provide an absolute path; if left empty, the following configuration will be used:
    ray_cluster_values:
      repository: "hub.oepkgs.net/oedeploy/rayproject/ray-with-pytorch" # Ray docker image URL
      tag: 2.9.0                       # Ray docker repository tag; when using the default docker source, x86 is 2.9.0, arm is 2.9.0-aarch64; please verify the specific tag in the configured repository.
      head:                            # Head node resources
        cpu: "1"
        memory: "2G"
      worker:                          # Worker node resources
        num: 1
        cpu: "1"
        memory: "1G"
    training:                          # Model training parameters
      pip: "https://pypi.tuna.tsinghua.edu.cn/simple"
      batch_size: 1024
      epoch: 5

Execute automated deployment:
Execute oedp run install in the plugin directory, or execute oedp run install -p [plugin directory] anywhere to complete the automated deployment of KubeRay.

View Dashboard

Query the corresponding port:

On the master node, use the kubectl get svc -A command to view port mappings:

bash

NAMESPACE     NAME                           TYPE        CLUSTER-IP      EXTERNAL-IP   PORT(S)                                                                       AGE
default       kubernetes                     ClusterIP   10.96.0.1       <none>        443/TCP                                                                       24h
kube-system   kube-dns                       ClusterIP   10.96.0.10      <none>        53/UDP,53/TCP,9153/TCP                                                        24h
kuberay       kuberay-operator               ClusterIP   10.96.175.101   <none>        8080/TCP                                                                      9h
kuberay       ray-cluster-kuberay-head-svc   NodePort    10.96.168.232   <none>        10001:32414/TCP,8265:31457/TCP,8080:31582/TCP,6379:31938/TCP,8000:32102/TCP   9h

The port corresponding to 8265 is the Dashboard port.

Open the Dashboard page: Use the http://ip:port/ link to open the Dashboard page, where you can view Ray Job / Serve / Cluster, resources, logs, and other information. The IP is the master node IP, and the port is the port corresponding to 8265 in step 3.1.

Demo 1: MLP Model Training and Inference Based on Dataset

KubeRay deployment is already completed. Pod resources require at least 4U.
On the master node, execute the following command to automatically complete model training and inference with one click.
bash
```
oedp run train -p kuberay
```

Demo 2: Time Comparison of Running the Same Training Model with Single Pod vs Multiple Pods

Before deploying KubeRay, configure Pod resources by modifying config.yaml.

Set the resources for a single Pod as:

bash

head:
  cpu: "7"
  memory: "14G"
worker:
  num: 0

Then execute:

bash

oedp run install -p kuberay
oedp run train -p kuberay

Set the resources for multiple Pods as:

bash

head:
  cpu: "7"
  memory: "14G"
worker:
  num: 2
  cpu: "5"
  memory: "10G"

Then execute:

bash

oedp run install -p kuberay
oedp run train -p kuberay

Compare the run times:
Under these conditions, the single Pod run time is about 25s, and the multiple Pods run time is about 67s.

Use Case 5: One-Click PyTorch Deployment

Prepare a k8s cluster.
Install the oeDeploy CLI tool oedp (default integrated in openEuler software sources released in 2025 and later).
bash
```
yum install -y oedp
```
Obtain and initialize the pytorch plugin.
shell
```
oedp repo update
oedp init pytorch
```
Modify config.yaml according to your actual situation. Ensure the target node is the k8s master node. kubectl_apply must correspond to the playbook under the workspace.

One-click deployment:

bash

oedp run install -p pytorch  # -p <plugin directory>

One-click uninstallation:

bash

oedp run uninstall -p pytorch  # -p <plugin directory>

Demo 1: Deploy and Print PyTorch Information

Modify config.yaml according to your actual situation. The following is an example:

yaml

all:
  hosts:
    localhost:
      ansible_connection: local
  vars:
    ansible_ssh_common_args: '-o StrictHostKeyChecking=no'
    # ================ demo 1: pytorch with http.server =====================
    kubectl_apply: pytorch-deployment.yaml
    namespace: pytorch-namespace
    replicas: 1
    containers:
      http:
        name: http-container
        image: hub.oepkgs.net/oedeploy/pytorch/pytorch:latest  # amd64
        # image: hub.oepkgs.net/oedeploy/pytorch/torchserve:latest-arm64  # arm64
      workspace_mount: /tmp
    service:
      port: 8080
      target_port: 8080
      node_port: 30699
    training:
      epoch: 2

Execute one-click deployment.
shell
```
oedp run install -p pytorch
```

View pods.

bash

kubectl get pods -n pytorch-namespace

text

NAME                                 READY   STATUS    RESTARTS   AGE
pytorch-deployment-db5d59bcb-ptqnp   1/1     Running   0          15m

View port mapping.

bash

kubectl get svc -n pytorch-namespace

text

NAME              TYPE       CLUSTER-IP     EXTERNAL-IP   PORT(S)        AGE
pytorch-service   NodePort   10.96.50.156   <none>        80:30699/TCP   15m

Access the web page.
text
```
http://x.x.x.x:30699/  # master node
```

Enter the container.

bash

kubectl exec -n pytorch-namespace -it pytorch-deployment-db5d59bcb-ptqnp -- /bin/bash

Print PyTorch information.

bash

python -c "import torch; print(torch.__version__); print(torch.tensor([1.0, 2.0, 3.0]) + torch.tensor([4.0, 5.0, 6.0]))"

Demo 2: Lightweight CNN Model Training Based on Dataset

Based on demo 1, PyTorch deployment is already completed.

Automatically complete model training with one click.

bash

oedp run train -p pytorch  # -p <plugin directory>

The output resembles:

bash

......
TASK [Display training output] *****************************************************************************************************************************************************************************************************************
ok: [localhost] => {
  "msg": [
    "Train Epoch: 0 [0/60000]\tLoss: 2.3114",
    "Train Epoch: 0 [6400/60000]\tLoss: 0.3884",
    "Train Epoch: 0 [12800/60000]\tLoss: 0.1483",
    "Train Epoch: 0 [19200/60000]\tLoss: 0.0510",
    "Train Epoch: 0 [25600/60000]\tLoss: 0.1151",
    "Train Epoch: 0 [32000/60000]\tLoss: 0.0191",
    "Train Epoch: 0 [38400/60000]\tLoss: 0.0690",
    "Train Epoch: 0 [44800/60000]\tLoss: 0.1995",
    "Train Epoch: 0 [51200/60000]\tLoss: 0.0417",
    "Train Epoch: 0 [57600/60000]\tLoss: 0.1821",
    "Test Accuracy: 9862/10000 (98.62%)",
    "Train Epoch: 1 [0/60000]\tLoss: 0.0052",
    ......
    "Train Epoch: 1 [57600/60000]\tLoss: 0.0115",
    "Test Accuracy: 9877/10000 (98.77%)"
  ]
}

Use Case 6: One-Click TensorFlow Deployment

Prepare a k8s cluster.
Install the oeDeploy CLI tool oedp (default integrated in openEuler software sources released in 2025 and later).
bash
```
yum install -y oedp
```
Obtain and initialize the tensorflow plugin.
shell
```
oedp repo update
oedp init tensorflow
```
Modify config.yaml according to your actual situation. Ensure the target node is the k8s master node. kubectl_apply must correspond to the playbook under the workspace.

One-click deployment:

bash

oedp run install -p tensorflow  # -p <plugin directory>

One-click uninstallation:

bash

oedp run uninstall -p tensorflow  # -p <plugin directory>

Demo 1: TensorFlow with Jupyter

Edit the config.yaml file.

yaml

all:
  hosts:
    localhost:
      ansible_connection: local
  
  vars:
    ansible_ssh_common_args: '-o StrictHostKeyChecking=no'

    kubectl_apply: tensorflow-deployment.yaml
    namespace: tensorflow-namespace
    replicas: 1
    containers:
      name: tensorflow-jupyter
      image: hub.oepkgs.net/oedeploy/tensorflow/tensorflow:latest-jupyter  # amd64 only
      command: ["jupyter", "notebook", "--ip=0.0.0.0", "--allow-root", "--no-browser"]
    service:
      name: tensorflow-service
      port: 80
      target_port: 8888
      node_port: 30088

View pod status.

bash

kubectl get pods -n tensorflow-namespace

text

NAME                                             READY   STATUS    RESTARTS   AGE
tensorflow-deployment-75b85948d8-w2n7b           1/1     Running   0          4m1s

Enter the pod container; there will be a prominent TensorFlow banner.

bash

kubectl exec -n tensorflow-namespace -it tensorflow-deployment-75b85948d8-w2n7b -- /bin/bash

Print TensorFlow information.

bash

python -c "import tensorflow as tf; print(tf.__version__)"

Open the Jupyter interface.
bash
```
http://x.x.x.x:30088/  # master node
```

Demo 2: Distributed Deployment Training Cluster

Edit the config.yaml file.

yaml

all:
  hosts:
    localhost:
      ansible_connection: local
  
  vars:
    ansible_ssh_common_args: '-o StrictHostKeyChecking=no'
    
    kubectl_apply: tensorflow-distributed.yaml
    namespace: tensorflow-namespace
    ps:
      replicas: 2
      containers:
        name: tensorflow-ps
        image: hub.oepkgs.net/oedeploy/tensorflow/tensorflow:latest  # amd64 only
      service:
        name: tensorflow-ps-service
        port: 2222
        target_port: 2222
    worker:
      replicas: 2
      containers:
        name: tensorflow-worker
        image: hub.oepkgs.net/oedeploy/tensorflow/tensorflow:latest  # amd64 only
      service:
        name: tensorflow-worker-service
        port: 2222
        target_port: 2222

View pod status.

bash

kubectl get pods -n tensorflow-namespace

text

NAME                                 READY   STATUS    RESTARTS   AGE
tensorflow-ps-fdddfdb5f-8fc98        1/1     Running   0          2m59s
tensorflow-ps-fdddfdb5f-dgqm9        1/1     Running   0          2m59s
tensorflow-worker-6cd8947b75-gxcnq   1/1     Running   0          2m59s
tensorflow-worker-6cd8947b75-wbt57   1/1     Running   0          2m59s

bash

kubectl logs tensorflow-ps-fdddfdb5f-8fc98 -n tensorflow-namespace

text

2025-02-07 09:25:24.882206: I tensorflow/core/util/port.cc:153] oneDNN custom operations are on. You may see slightly different numerical results due to floating-point round-off errors from different computation orders. To turn them off, set the environment variable `TF_ENABLE_ONEDNN_OPTS=0`.
2025-02-07 09:25:24.915813: I tensorflow/core/platform/cpu_feature_guard.cc:210] This TensorFlow binary is optimized to use available CPU instructions in performance-critical operations.
To enable the following instructions: AVX2 AVX512F AVX512_VNNI FMA, in other operations, rebuild TensorFlow with the appropriate compiler flags.
TensorFlow version: 2.18.0
Environment variables: environ({'PATH': '/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin', 'HOSTNAME': 'tensorflow-ps-fdddfdb5f-8fc98', 'DEBIAN_FRONTEND': 'noninteractive', 'LANG': 'C.UTF-8', 'TF_CONFIG': '{\n  "cluster": {\n    "ps": ["tensorflow-ps:2222"],\n    "worker": ["tensorflow-worker-0:2222", "tensorflow-worker-1:2222"]\n  },\n  "task": {"type": "ps", "index": 0}\n}\n', 'KUBERNETES_PORT_443_TCP_PROTO': 'tcp', 'TENSORFLOW_PS_PORT_2222_TCP_PROTO': 'tcp', 'TENSORFLOW_WORKER_SERVICE_PORT': '2222', 'TENSORFLOW_WORKER_PORT': 'tcp://10.96.56.106:2222', 'KUBERNETES_SERVICE_HOST': '10.96.0.1', 'KUBERNETES_PORT_443_TCP_PORT': '443', 'TENSORFLOW_PS_PORT_2222_TCP_PORT': '2222', 'TENSORFLOW_WORKER_PORT_2222_TCP_ADDR': '10.96.56.106', 'TENSORFLOW_WORKER_SERVICE_HOST': '10.96.56.106', 'TENSORFLOW_WORKER_PORT_2222_TCP_PROTO': 'tcp', 'KUBERNETES_SERVICE_PORT': '443', 'KUBERNETES_SERVICE_PORT_HTTPS': '443', 'KUBERNETES_PORT': 'tcp://10.96.0.1:443', 'KUBERNETES_PORT_443_TCP': 'tcp://10.96.0.1:443', 'TENSORFLOW_PS_PORT_2222_TCP': 'tcp://10.96.88.2:2222', 'TENSORFLOW_PS_PORT_2222_TCP_ADDR': '10.96.88.2', 'TENSORFLOW_WORKER_PORT_2222_TCP_PORT': '2222', 'KUBERNETES_PORT_443_TCP_ADDR': '10.96.0.1', 'TENSORFLOW_PS_SERVICE_HOST': '10.96.88.2', 'TENSORFLOW_PS_SERVICE_PORT': '2222', 'TENSORFLOW_PS_PORT': 'tcp://10.96.88.2:2222', 'TENSORFLOW_WORKER_PORT_2222_TCP': 'tcp://10.96.56.106:2222', 'HOME': '/root', 'ENABLE_RUNTIME_UPTIME_TELEMETRY': '1', 'TF2_BEHAVIOR': '1', 'TPU_ML_PLATFORM': 'Tensorflow', 'TPU_ML_PLATFORM_VERSION': '2.18.0'})
This is the parameter server.
[2025-02-07 09:25:26.288107] Parameter server is running...
[2025-02-07 09:26:26.288219] Parameter server is running...
[2025-02-07 09:27:26.288358] Parameter server is running...
[2025-02-07 09:28:26.288519] Parameter server is running...

Uninstall pods.

bash

kubectl delete -f tensorflow-distributed.yaml

oeDeploy Use Cases ​

Use Case 1: One-Click Kubernetes Deployment ​

Use Case 2: One-Click AnythingLLM Deployment ​

Demo: Deploy and Print AnythingLLM Information ​

Use Case 3: One-Click DeepSeek-R1 Deployment ​

Use Case 4: One-Click KubeRay Deployment ​

Environment Preparation ​

Deploy KubeRay ​

View Dashboard ​

Demo 1: MLP Model Training and Inference Based on Dataset ​

Demo 2: Time Comparison of Running the Same Training Model with Single Pod vs Multiple Pods ​

Use Case 5: One-Click PyTorch Deployment ​

Demo 1: Deploy and Print PyTorch Information ​

Demo 2: Lightweight CNN Model Training Based on Dataset ​

Use Case 6: One-Click TensorFlow Deployment ​

Demo 1: TensorFlow with Jupyter ​

Demo 2: Distributed Deployment Training Cluster ​

oeDeploy Use Cases

Use Case 1: One-Click Kubernetes Deployment

Use Case 2: One-Click AnythingLLM Deployment

Demo: Deploy and Print AnythingLLM Information

Use Case 3: One-Click DeepSeek-R1 Deployment

Use Case 4: One-Click KubeRay Deployment

Environment Preparation

Deploy KubeRay

View Dashboard

Demo 1: MLP Model Training and Inference Based on Dataset

Demo 2: Time Comparison of Running the Same Training Model with Single Pod vs Multiple Pods

Use Case 5: One-Click PyTorch Deployment

Demo 1: Deploy and Print PyTorch Information

Demo 2: Lightweight CNN Model Training Based on Dataset

Use Case 6: One-Click TensorFlow Deployment

Demo 1: TensorFlow with Jupyter

Demo 2: Distributed Deployment Training Cluster