tutorial_01.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Miniminiai Tutorial 1: Introduction\n",
    "\n",
    "This is a minimal example to get you started, showing the basic flow of training a model using (mini)miniai.\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Setup\n",
    "\n",
    "Installing the library and importing a few useful things"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Imports\n",
    "import torch.nn as nn\n",
    "import minai as mi # So we can see what is from minai in this tutorial\n",
    "import torchvision.transforms.functional as TF\n",
    "from datasets import load_dataset\n",
    "from torcheval.metrics import  MulticlassAccuracy"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Preparing the DataLoaders\n",
    "\n",
    "The dataloaders is just a tiny wrapper around two pytorch dataloaders, dls.train and dls.valid. You can create your dataloaders with `dls=DataLoaders(train_dl, valid_dl)` or use the `from_dd` method like we do here to load them from a DatasetDict (for datasets from huggingface with the datasets library):\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:datasets.builder:Found cached dataset mnist (/root/.cache/huggingface/datasets/mnist/mnist/1.0.0/fda16c03c4ecfb13f165ba7e29cf38129ce035011519968cdaf74894ce91c9d4)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "65be2a364ec8486a9bd46d20deed89a1",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/2 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "(torch.Size([64, 1, 28, 28]), torch.Size([64]), tensor([8, 4, 0, 3, 6]))"
      ]
     },
     "execution_count": null,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Load a dataset from HF\n",
    "dataset = load_dataset('mnist')\n",
    "\n",
    "# Specify transforms\n",
    "def transforms(b):\n",
    "    b['image'] = [TF.to_tensor(o) for o in b['image']]\n",
    "    return b\n",
    "dataset = dataset.with_transform(transforms)\n",
    "\n",
    "# Turn it into dls\n",
    "dls = mi.DataLoaders.from_dd(dataset,batch_size=64)\n",
    "\n",
    "# Look at the data\n",
    "xb, yb = next(iter(dls.train))\n",
    "xb.shape, yb.shape, yb[:5]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1500x300 with 5 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# The library has some useful utility functions such as:\n",
    "mi.show_images(xb[:5], ncols=5, titles=list(yb[:5].numpy()))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "You can do a lot of fancy stuff with your collate function if your data requires more processing or augmentation."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Prepare the Model\n",
    "\n",
    "The model can be pretty much any PyTorch model, no changes needed here :)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = nn.Sequential(\n",
    "    nn.Conv2d(1, 16, kernel_size=3, stride=2, padding=1),\n",
    "    nn.ReLU(),\n",
    "    nn.Conv2d(16, 16, kernel_size=3, stride=2, padding=1),\n",
    "    nn.ReLU(),\n",
    "    nn.Conv2d(16, 10, kernel_size=3, stride=2, padding=1),\n",
    "    nn.ReLU(),\n",
    "    nn.AdaptiveAvgPool2d(1),\n",
    "    nn.Flatten()\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Create and Fit the Learner\n",
    "\n",
    "The heart of (mini)miniai is the Learner class. It pulls together the data, model and loss function, and can be extended in all sorts of cool ways using callbacks. Here's a somewhat minimal example, training our model on this classification task and plotting some stats as we do so:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "\n",
       "<style>\n",
       "    /* Turns off some styling */\n",
       "    progress {\n",
       "        /* gets rid of default border in Firefox and Opera. */\n",
       "        border: none;\n",
       "        /* Needs to be in here for Safari polyfill so background images work as expected. */\n",
       "        background-size: auto;\n",
       "    }\n",
       "    progress:not([value]), progress:not([value])::-webkit-progress-bar {\n",
       "        background: repeating-linear-gradient(45deg, #7e7e7e, #7e7e7e 10px, #5c5c5c 10px, #5c5c5c 20px);\n",
       "    }\n",
       "    .progress-bar-interrupted, .progress-bar-interrupted::-webkit-progress-bar {\n",
       "        background: #F44336;\n",
       "    }\n",
       "</style>\n"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: left;\">\n",
       "      <th>accuracy</th>\n",
       "      <th>loss</th>\n",
       "      <th>epoch</th>\n",
       "      <th>train</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>0.337</td>\n",
       "      <td>1.853</td>\n",
       "      <td>0</td>\n",
       "      <td>train</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>0.587</td>\n",
       "      <td>1.261</td>\n",
       "      <td>0</td>\n",
       "      <td>eval</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>0.707</td>\n",
       "      <td>0.906</td>\n",
       "      <td>1</td>\n",
       "      <td>train</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>0.801</td>\n",
       "      <td>0.648</td>\n",
       "      <td>1</td>\n",
       "      <td>eval</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>0.819</td>\n",
       "      <td>0.586</td>\n",
       "      <td>2</td>\n",
       "      <td>train</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>0.838</td>\n",
       "      <td>0.522</td>\n",
       "      <td>2</td>\n",
       "      <td>eval</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# There are callbacks for all sorts of things, here are some common ones:\n",
    "cbs = [\n",
    "    mi.TrainCB(), # Handles the core steps in the training loop. Can be left out if using TrainLearner\n",
    "    mi.DeviceCB(), # Handles making sure data and model are on the right device\n",
    "    mi.MetricsCB(accuracy=MulticlassAccuracy()), # Keep track of any relevant metrics\n",
    "    mi.ProgressCB(), # Displays metrics and loss during training, optionally plot=True for a pretty graph\n",
    "]\n",
    "\n",
    "# Nothing fancy for the loss function\n",
    "loss_fn = nn.CrossEntropyLoss()\n",
    "\n",
    "# The learner takes a model, dataloaders and loss function, plus some optional extras like a list of callbacks\n",
    "learn = mi.Learner(model, dls, loss_fn, lr=0.1, cbs=cbs)\n",
    "\n",
    "# And fit does the magic :)\n",
    "learn.fit(3)"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "When I get around to making more complex tutorials I'll try to show some of the other existing callbacks in action, but for most tasks this is pretty much all you need! The model (`learn.model`) is just a regular PyTorch model, so you can save it and load it later somewhere else without needing any minai code at all. "
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "python3",
   "language": "python",
   "name": "python3"
  },
  "widgets": {
   "application/vnd.jupyter.widget-state+json": {
    "243a6081666e4c1497d740e8c895b52a": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "2e415547b49048fb8a7fd9c0cf93fd8c": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "2fa7aece2e854579b6856dc227d90383": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "HTMLModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "HTMLModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "HTMLView",
      "description": "",
      "description_tooltip": null,
      "layout": "IPY_MODEL_f19e76960da346debdde6e0b89f291cf",
      "placeholder": "​",
      "style": "IPY_MODEL_b94f987ac374487da181bf3c73d3dec4",
      "value": " 2/2 [00:00&lt;00:00, 92.52it/s]"
     }
    },
    "44e64afc3b3f49f9b81b05a88fc924c8": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "ProgressStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "ProgressStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "bar_color": null,
      "description_width": ""
     }
    },
    "65be2a364ec8486a9bd46d20deed89a1": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "HBoxModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "HBoxModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "HBoxView",
      "box_style": "",
      "children": [
       "IPY_MODEL_862e35aad3494ea1837743ae186ec3b9",
       "IPY_MODEL_a3c7c7de120549ecaabff8615a6ecc10",
       "IPY_MODEL_2fa7aece2e854579b6856dc227d90383"
      ],
      "layout": "IPY_MODEL_243a6081666e4c1497d740e8c895b52a"
     }
    },
    "862e35aad3494ea1837743ae186ec3b9": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "HTMLModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "HTMLModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "HTMLView",
      "description": "",
      "description_tooltip": null,
      "layout": "IPY_MODEL_2e415547b49048fb8a7fd9c0cf93fd8c",
      "placeholder": "​",
      "style": "IPY_MODEL_eec16c8f02f147c98ce4d21f40f3782f",
      "value": "100%"
     }
    },
    "a3c7c7de120549ecaabff8615a6ecc10": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "FloatProgressModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "FloatProgressModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "ProgressView",
      "bar_style": "success",
      "description": "",
      "description_tooltip": null,
      "layout": "IPY_MODEL_afd0138634664df38f89b59f6949970a",
      "max": 2,
      "min": 0,
      "orientation": "horizontal",
      "style": "IPY_MODEL_44e64afc3b3f49f9b81b05a88fc924c8",
      "value": 2
     }
    },
    "afd0138634664df38f89b59f6949970a": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "b94f987ac374487da181bf3c73d3dec4": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "DescriptionStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "DescriptionStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "description_width": ""
     }
    },
    "eec16c8f02f147c98ce4d21f40f3782f": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "DescriptionStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "DescriptionStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "description_width": ""
     }
    },
    "f19e76960da346debdde6e0b89f291cf": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    }
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}