{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "[![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/duoan/TorchCode/blob/master/templates/27_vit_patch.ipynb)\n",
    "\n",
    "# 🟠 Medium: ViT Patch Embedding\n",
    "\n",
    "Implement the **patch embedding** layer from Vision Transformer (ViT).\n",
    "\n",
    "### Signature\n",
    "```python\n",
    "class PatchEmbedding(nn.Module):\n",
    "    def __init__(self, img_size, patch_size, in_channels, embed_dim): ...\n",
    "    def forward(self, x: Tensor) -> Tensor:\n",
    "        # x: (B, C, H, W)\n",
    "        # Returns: (B, num_patches, embed_dim)\n",
    "```\n",
    "\n",
    "### Algorithm\n",
    "1. Reshape image into non-overlapping patches: `(B, C, H, W)` → `(B, N, C*P*P)`\n",
    "2. Project each patch: `nn.Linear(C*P*P, embed_dim)`\n",
    "3. `num_patches = (img_size // patch_size) ** 2`"
   ],
   "outputs": []
  },
  {
   "cell_type": "code",
   "metadata": {},
   "source": [
    "# Install torch-judge in Colab (no-op in JupyterLab/Docker)\n",
    "try:\n",
    "    import google.colab\n",
    "    get_ipython().run_line_magic('pip', 'install -q torch-judge')\n",
    "except ImportError:\n",
    "    pass\n"
   ],
   "outputs": [],
   "execution_count": null
  },
  {
   "cell_type": "code",
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch\n",
    "import torch.nn as nn"
   ],
   "execution_count": null
  },
  {
   "cell_type": "code",
   "metadata": {},
   "outputs": [],
   "source": [
    "# ✏️ YOUR IMPLEMENTATION HERE\n",
    "\n",
    "class PatchEmbedding(nn.Module):\n",
    "    def __init__(self, img_size, patch_size, in_channels, embed_dim):\n",
    "        super().__init__()\n",
    "        pass  # self.num_patches, self.proj\n",
    "\n",
    "    def forward(self, x):\n",
    "        pass  # reshape to patches, project"
   ],
   "execution_count": null
  },
  {
   "cell_type": "code",
   "metadata": {},
   "outputs": [],
   "source": [
    "# 🧪 Debug\n",
    "pe = PatchEmbedding(32, 8, 3, 64)\n",
    "x = torch.randn(2, 3, 32, 32)\n",
    "print('Output:', pe(x).shape)\n",
    "print('Patches:', pe.num_patches)"
   ],
   "execution_count": null
  },
  {
   "cell_type": "code",
   "metadata": {},
   "outputs": [],
   "source": [
    "# ✅ SUBMIT\n",
    "from torch_judge import check\n",
    "check('vit_patch')"
   ],
   "execution_count": null
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "name": "python",
   "version": "3.11.0"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}