vessl-intro.ipynb

{
 "cells": [
  {
   "attachments": {
    "image-4.png": {
     "image/png": "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"
    }
   },
   "cell_type": "markdown",
   "id": "dc6f90c0",
   "metadata": {},
   "source": [
    "Vessl - Using Python SDK on Jupyter Notebook\n",
    "------\n",
    "\n",
    "![image-4.png](attachment:image-4.png)\n",
    "\n",
    "[Vessl](https://vessl.ai)은 머신러닝 연구자들이 팀과 함께 모델 학습과 모니터링을 더 효과적으로 할 수 있는 머신러닝 플랫폼입니다. 이 노트북에서는 유저의 로컬 환경에서 수행한 실험을 어떻게 Vessl과 연동하여 더 쉽고 효율적으로 연구를 할 수 있는지 설명드릴 예정입니다. Vessl에 대한 더 자세한 설명이 궁금하시면, [공식 문서](https://docs.vessl.ai/)를 참고해주세요!\n",
    "\n",
    "[Vessl](https://vessl.ai) is a machine learning platform that allows researchers to train and monitor models in scale as a team. In this notebook, you'll see how you can integrate local ML experiments with Vessl to empower your research workflow. For more detailed information, check out our [docs](https://docs.vessl.ai/)!"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "40ac4820",
   "metadata": {},
   "source": [
    "### 1. Install Vessl SDK and required libraries\n",
    "\n",
    "먼저 Vessl SDK와 간단한 실험에 필요한 라이브러리들을 설치해보겠습니다. 지금 Vessl의 workspace 환경에서 이 노트북을 실행하고 계시다면 Vessl SDK를 포함해서 대부분의 라이브러리들이 설치되어 있을 것입니다. 이 노트북에서의 예제는 Vessl SDK 0.1.15 버전 이상에서 작동합니다.\n",
    "\n",
    "Let's install the Vessl SDK and required libraries first. If you're running this Jupyter Notebook on a Vessl workspace, most of these should already be installed. Examples in this Notebook run well on Vessl SDK versions 0.1.15 or higher."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "f733673b",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Required libraries\n",
    "!!pip install matplotlib hypothesis torch torchvision watermark hypothesis"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "b7eda1d0",
   "metadata": {},
   "outputs": [],
   "source": [
    "# This cell should return: ['vessl, version 0.1.15']\n",
    "!!pip install --upgrade vessl\n",
    "!!vessl --version"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "7d3a557c",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Use the CLI to sign in into Vessl and configure a default organization and project\n",
    "!!vessl configure"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ce2dfa90",
   "metadata": {},
   "outputs": [],
   "source": [
    "# You can also programatically set a default organization and project using the Python API\n",
    "import vessl\n",
    "vessl.configure(organization_name=\"vessl\", project_name=\"examples\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "a4a5ed24",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Check other dependencies\n",
    "import os\n",
    "import matplotlib\n",
    "import matplotlib.pyplot as plt\n",
    "import hypothesis\n",
    "import torch\n",
    "from torch import optim\n",
    "import torch.nn.functional as F\n",
    "from torch.optim.lr_scheduler import StepLR\n",
    "import watermark\n",
    "\n",
    "import vessl \n",
    "\n",
    "%pylab\n",
    "%matplotlib inline\n",
    "%load_ext watermark\n",
    "%watermark -v -p numpy,torch,torchvision"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d9bff975",
   "metadata": {},
   "source": [
    "### 2. Prepare model code and dataset\n",
    "\n",
    "이 예제에서는 간단한 2D CNN 기반 인공 신경망을 [MNIST](https://paperswithcode.com/dataset/mnist) 데이터셋으로 학습하여 0에서 9까지의 숫자를 구분하는 머신러닝 모델을 만들어보겠습니다. 예제에서는 공개되어 있는 torchvision의 데이터셋을 사용하였으나, Vessl의 [Dataset](https://docs.vessl.ai/user-guide/dataset) 기능을 이용하여 필요한 어떤 데이터셋이든 Vessl에서 불러와 사용하실 수 있습니다!\n",
    "\n",
    "In this example, we will train an image classification model classifying numeric digits (0-9) using a simple 2D CNN and the [MNIST](https://paperswithcode.com/dataset/mnist) dataset. We used a publicly available dataset from torchvision for this example, but you can also import you own datasets using the [Dataset](https://docs.vessl.ai/user-guide/dataset) feature in Vessl!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "7b6abf54",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Download MNIST training data\n",
    "from torchvision import datasets\n",
    "from torchvision.transforms import ToTensor\n",
    "train_data = datasets.MNIST(\n",
    "    root = 'data',\n",
    "    train = True,                         \n",
    "    transform = ToTensor(), \n",
    "    download = True,            \n",
    ")\n",
    "test_data = datasets.MNIST(\n",
    "    root = 'data', \n",
    "    train = False, \n",
    "    transform = ToTensor()\n",
    ")\n",
    "\n",
    "print(f'The shape of train data: {train_data.data.shape}')\n",
    "print(f'The shape of test data: {test_data.data.shape}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "72518ad7",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Render some sample images from training data\n",
    "figure = plt.figure(figsize=(10, 8))\n",
    "cols, rows = 5, 5\n",
    "\n",
    "for i in range(1, cols*rows + 1):\n",
    "    sample_idx = torch.randint(len(train_data), size=(1,)).item()\n",
    "    img, label = train_data[sample_idx]\n",
    "    figure.add_subplot(rows, cols, i)\n",
    "    plt.title(label)\n",
    "    plt.axis(\"off\")\n",
    "    plt.imshow(img.squeeze(), cmap=\"gray\")\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "90e127d4",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Define data loader\n",
    "from torch.utils.data import DataLoader\n",
    "\n",
    "train_loader = torch.utils.data.DataLoader(train_data, batch_size=128, shuffle=True, num_workers=1)\n",
    "test_loader = torch.utils.data.DataLoader(test_data, batch_size=128, shuffle=False, num_workers=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "d9b66f60",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Define model\n",
    "import torch.nn as nn\n",
    "\n",
    "class CNN(nn.Module):\n",
    "    def __init__(self):\n",
    "        super(CNN, self).__init__()\n",
    "        self.conv1 = nn.Conv2d(1, 32, 3, 1)\n",
    "        self.conv2 = nn.Conv2d(32, 64, 3, 1)\n",
    "        self.relu = nn.ReLU()\n",
    "        self.pool = nn.MaxPool2d(2)\n",
    "        self.drop1 = nn.Dropout2d(0.25)\n",
    "        self.drop2 = nn.Dropout2d(0.5)\n",
    "        self.fc1 = nn.Linear(9216, 128)\n",
    "        self.fc2 = nn.Linear(128, 10)\n",
    "        self.softmax = nn.LogSoftmax(1)\n",
    "\n",
    "    def forward(self, x):\n",
    "        x = self.relu(self.conv1(x))\n",
    "        x = self.relu(self.conv2(x))\n",
    "        x = self.pool(x)\n",
    "        x = self.drop1(x)\n",
    "        x = torch.flatten(x, 1)\n",
    "        x = self.fc1(x)\n",
    "        x = self.relu(x)\n",
    "        x = self.drop2(x)\n",
    "        x = self.fc2(x)\n",
    "        return self.softmax(x)\n",
    "    \n",
    "# Initialize model\n",
    "device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')\n",
    "model = CNN().to(device)\n",
    "if torch.cuda.device_count() > 1:\n",
    "    model = nn.DataParallel(model)\n",
    "\n",
    "print(model)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2b8db035",
   "metadata": {},
   "source": [
    "### 3. Use vessl.log() to record model training and test steps\n",
    "\n",
    "모델이 준비되었다면 트레이닝을 위한 스크립트를 작성할 차례입니다. Vessl에서는 vessl.log() 라는 간단한 인터페이스를 통해 실험의 메트릭과 로그, 추론 결과물 등 실험의 다양한 출력들을 기록할 수 있습니다. 기록한 결과물들은 Vessl의 실험 대시보드를 통해 언제든 확인하고, 다른 실험과 결과물들을 비교하고, 더 나은 결과를 낸 실험을 모델로 저장해둘 수 있습니다. 여러분의 학습 코드에서 기록이 필요한 곳에 vessl.log()를 넣어두기만 하면 됩니다!\n",
    "* vessl.log() 에 대한 자세한 내용은 [공식 문서의 SDK 항목](https://docs.vessl.ai/api-reference/python-sdk/vessl.log)을 참고해주세요.\n",
    "* 실험 결과 기록과 조회에 대한 자세한 내용은 [공식 문서의 Experiment 항목](https://docs.vessl.ai/user-guide/experiment)을 참고해주세요.\n",
    "\n",
    "When the model code is ready, the next step is to write a script for training and testing the model. With Vessl, you can record outputs from your training such as metrics, logs, and inferences. You can always view the results, compare with other experiments, and save the best one as a model on our Experiment Tracking page. The only thing you need to do is add vessl.log() in your training code!\n",
    "* For more information about vessl.log(), check out our [Python SDK docs](https://docs.vessl.ai/api-reference/python-sdk/vessl.log).\n",
    "* For more information on experiment tracking, check out our [experiment docs](https://docs.vessl.ai/user-guide/experiment)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "4e802f48",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Train function (with vessl logging)\n",
    "from torch.autograd import Variable\n",
    "\n",
    "def train(model, device, train_loader, optimizer, epoch, start_epoch):\n",
    "    model.train()\n",
    "    loss = 0\n",
    "    for batch_idx, (data, label) in enumerate(train_loader):\n",
    "        data, label = data.to(device), label.to(device)\n",
    "        optimizer.zero_grad()\n",
    "        output = model(data)\n",
    "        loss = F.nll_loss(output, label)\n",
    "        loss.backward()\n",
    "        optimizer.step()\n",
    "        if batch_idx % 128 == 0:\n",
    "            print('Train Epoch: {} [{}/{} ({:.0f}%)]\\tLoss: {:.6f}'.format(\n",
    "                epoch + 1, batch_idx * len(data), len(train_loader.dataset),\n",
    "                100. * batch_idx / len(train_loader), loss.item()))\n",
    "\n",
    "    # Logging loss metrics to Vessl\n",
    "    vessl.log(\n",
    "        step=epoch + start_epoch + 1,\n",
    "        payload={'loss': loss.item()}\n",
    "    )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "abc3a3e1",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Test function (with vessl plot uploading)\n",
    "def test(model, device, test_loader, save_image):\n",
    "    model.eval()\n",
    "    test_loss = 0\n",
    "    correct = 0\n",
    "    test_images = []\n",
    "    with torch.no_grad():\n",
    "        for data, target in test_loader:\n",
    "            data, target = data.to(device), target.to(device)\n",
    "            output = model(data)\n",
    "            test_loss += F.nll_loss(output, target, reduction='sum').item()\n",
    "            pred = output.argmax(dim=1, keepdim=True)\n",
    "            correct += pred.eq(target.view_as(pred)).sum().item()\n",
    "            test_images.append(vessl.Image(\n",
    "                data[0], caption=\"Pred: {} Truth: {}\".format(pred[0].item(), target[0])))\n",
    "\n",
    "    test_loss /= len(test_loader.dataset)\n",
    "    test_accuracy = 100. * correct / len(test_loader.dataset)\n",
    "\n",
    "    print('\\nTest set: Average loss: {:.4f}, Accuracy: {}/{} ({:.0f}%)\\n'.format(\n",
    "        test_loss, correct, len(test_loader.dataset), test_accuracy))\n",
    "\n",
    "    # Saving inference results with caption into Vessl\n",
    "    if save_image:\n",
    "        vessl.log({\"Examples\": test_images})\n",
    "\n",
    "    return test_accuracy"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "55a6fa94",
   "metadata": {},
   "source": [
    "### 4. Train the model\n",
    "\n",
    "모든 준비가 완료되었으면 실험을 수행해봅시다! vessl.init() 을 이용해서 새로운 실험을 생성 후 실험 스크립트를 실행하면 자동으로 생성된 실험에 실험 기록들이 저장됩니다. 노트북 환경에서는 실험이 완료된 후 vessl.finish()로 실험을 종료할 수 있습니다. 노트북이 아닌 실제 실험 환경에서는 스크립트를 실행하는 프로세스가 종료될 시 자동으로 vessl.finish()가 호출되므로 직접 호출을 하실 필요가 없습니다.\n",
    "\n",
    "Let's execute the actual experiment! You can create a new experiment by calling vessl.init() and our SDK will record all metrics and logs onto the newly created experiment. After the experiment is finished, you can call vessl.finish() to mark experiment as completed. We automatically call vessl.finish() when your script exits so the last step is optional."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "d28c0fa1",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Initialize new experiment via Vessl SDK \n",
    "vessl.init()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "380a82b2",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Hyperparameters\n",
    "epochs = 10\n",
    "batch_size = 128\n",
    "learning_rate = 0.01\n",
    "optimizer = optim.Adam(model.parameters(), lr=learning_rate)\n",
    "scheduler = StepLR(optimizer, step_size=1, gamma=0.7)\n",
    "\n",
    "# Train the model\n",
    "for epoch in range(epochs):\n",
    "    train(model, device, train_loader, optimizer, epoch, 0)\n",
    "    test(model, device, test_loader, True)\n",
    "\n",
    "    scheduler.step()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "fd1f9626",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Finish experiment\n",
    "vessl.finish()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4315c5cd",
   "metadata": {},
   "source": [
    "### 5. More Informations\n",
    "\n",
    "Python SDK을 이용한 실험 기록/관리 기능 외에도, Vessl은 더 효율적인 머신러닝 연구를 위한 다양한 기능을 지원합니다! 아래와 같은 기능들이 있습니다.\n",
    "- 로컬 환경 대신 Vessl 클러스터에서 실험을 실행하여, 여러개의 실험이 동시에 수행되는 동안 다른 중요한 일들에 집중하실 수 있습니다. [Creating an Experiment](https://docs.vessl.ai/user-guide/experiment/creating-an-experiment) 항목을 참고해주세요.\n",
    "- 대규모의 컴퓨팅 자원이 필요할 경우 여러 노드에서 분산 학습을 진행하실 수 있습니다. [Distributed Experiment](https://docs.vessl.ai/user-guide/experiment/distributed-experiment) 항목을 참고해주세요.\n",
    "- 최적의 결과를 내기 위한 하이퍼파라미터를 자동으로 탐색해주는 Hyperparamter Optimization 기능을 지원합니다. [Sweep](https://docs.vessl.ai/user-guide/sweep) 항목을 참고해주세요.\n",
    "\n",
    "그 외에도 Vessl의 다양한 기능들을 이용하여 머신러닝 작업의 생산성을 극대화하실 수 있습니다! [커뮤니티 슬랙에 가입](https://join.slack.com/t/vessl-ai-community/shared_invite/zt-w7phk2pj-XwGK42BchBv5qeXYH42iNQ)하셔서 다른 머신러닝 연구자 및 엔지니어들과 교류하고, Vessl 팀과 질의응답을 하시거나 Best practice들을 서로 공유하실 수 있습니다. 감사합니다!\n",
    "\n",
    "\n",
    "Vessl supports many other features that will improve your machine learning research workflow! For instance:\n",
    "- You can also execute your training job on top of Vessl-managed clusters to run the experiments asynchronously and focus on other important things. See [Creating an Experiment](https://docs.vessl.ai/user-guide/experiment/creating-an-experiment) for more information.\n",
    "- Vessl supports distributed experiments for training that needs large amounts of computing resources - See [Distributed Experiment](https://docs.vessl.ai/user-guide/experiment/distributed-experiment) for more information.\n",
    "- Hyperparameter optimzation can also be done easily with Vessl! See [Sweep](https://docs.vessl.ai/user-guide/sweep) for more information.\n",
    "\n",
    "Maximize your research productivity with Vessl - Join our [Slack community](https://join.slack.com/t/vessl-ai-community/shared_invite/zt-w7phk2pj-XwGK42BchBv5qeXYH42iNQ) to chat with other ML researchers, ask questions, provide feedback, or share best practices. Thank you very much!"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.9.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}