trace_graph.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# brew install graphviz\n",
    "# pip install graphviz\n",
    "from graphviz import Digraph"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from micrograd.engine import Value"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "def trace(root):\n",
    "    nodes, edges = set(), set()\n",
    "    def build(v):\n",
    "        if v not in nodes:\n",
    "            nodes.add(v)\n",
    "            for child in v._prev:\n",
    "                edges.add((child, v))\n",
    "                build(child)\n",
    "    build(root)\n",
    "    return nodes, edges\n",
    "\n",
    "def draw_dot(root, format='svg', rankdir='LR'):\n",
    "    \"\"\"\n",
    "    format: png | svg | ...\n",
    "    rankdir: TB (top to bottom graph) | LR (left to right)\n",
    "    \"\"\"\n",
    "    assert rankdir in ['LR', 'TB']\n",
    "    nodes, edges = trace(root)\n",
    "    dot = Digraph(format=format, graph_attr={'rankdir': rankdir}) #, node_attr={'rankdir': 'TB'})\n",
    "    \n",
    "    for n in nodes:\n",
    "        dot.node(name=str(id(n)), label = \"{ data %.4f | grad %.4f }\" % (n.data, n.grad), shape='record')\n",
    "        if n._op:\n",
    "            dot.node(name=str(id(n)) + n._op, label=n._op)\n",
    "            dot.edge(str(id(n)) + n._op, str(id(n)))\n",
    "    \n",
    "    for n1, n2 in edges:\n",
    "        dot.edge(str(id(n1)), str(id(n2)) + n2._op)\n",
    "    \n",
    "    return dot"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# a very simple example\n",
    "x = Value(1.0)\n",
    "y = (x * 2 + 1).relu()\n",
    "y.backward()\n",
    "draw_dot(y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# a simple 2D neuron\n",
    "import random\n",
    "from micrograd import nn\n",
    "\n",
    "random.seed(1337)\n",
    "n = nn.Neuron(2)\n",
    "x = [Value(1.0), Value(-2.0)]\n",
    "y = n(x)\n",
    "y.backward()\n",
    "\n",
    "dot = draw_dot(y)\n",
    "dot"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "dot.render('gout')"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "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": "3.6.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}