chapter 3

chapter3.ipynb

@@ -0,0 +1,428 @@
+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "\n",
+    "from fastai.vision.all import *\n",
+    "import gradio as gr\n",
+    "import timm"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 3,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "['convnext_atto',\n",
+       " 'convnext_atto_ols',\n",
+       " 'convnext_base',\n",
+       " 'convnext_base_384_in22ft1k',\n",
+       " 'convnext_base_in22ft1k',\n",
+       " 'convnext_base_in22k',\n",
+       " 'convnext_femto',\n",
+       " 'convnext_femto_ols',\n",
+       " 'convnext_large',\n",
+       " 'convnext_large_384_in22ft1k',\n",
+       " 'convnext_large_in22ft1k',\n",
+       " 'convnext_large_in22k',\n",
+       " 'convnext_nano',\n",
+       " 'convnext_nano_ols',\n",
+       " 'convnext_pico',\n",
+       " 'convnext_pico_ols',\n",
+       " 'convnext_small',\n",
+       " 'convnext_small_384_in22ft1k',\n",
+       " 'convnext_small_in22ft1k',\n",
+       " 'convnext_small_in22k',\n",
+       " 'convnext_tiny',\n",
+       " 'convnext_tiny_384_in22ft1k',\n",
+       " 'convnext_tiny_hnf',\n",
+       " 'convnext_tiny_in22ft1k',\n",
+       " 'convnext_tiny_in22k',\n",
+       " 'convnext_xlarge_384_in22ft1k',\n",
+       " 'convnext_xlarge_in22ft1k',\n",
+       " 'convnext_xlarge_in22k']"
+      ]
+     },
+     "execution_count": 3,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "timm.list_models('convnext*')"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 4,
+   "metadata": {},
+   "outputs": [
+    {
+     "ename": "NameError",
+     "evalue": "name 'dls' is not defined",
+     "output_type": "error",
+     "traceback": [
+      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
+      "\u001b[1;31mNameError\u001b[0m                                 Traceback (most recent call last)",
+      "Cell \u001b[1;32mIn[4], line 1\u001b[0m\n\u001b[1;32m----> 1\u001b[0m learn \u001b[39m=\u001b[39m vision_learner(dls , \u001b[39m'\u001b[39m\u001b[39mconvnext_tiny_in22k\u001b[39m\u001b[39m'\u001b[39m , metrics \u001b[39m=\u001b[39m error_rate)\u001b[39m.\u001b[39mto_fp16()\n\u001b[0;32m      2\u001b[0m learn\u001b[39m.\u001b[39mfine_tune(\u001b[39m3\u001b[39m)\n",
+      "\u001b[1;31mNameError\u001b[0m: name 'dls' is not defined"
+     ]
+    }
+   ],
+   "source": []
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 6,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "10.75"
+      ]
+     },
+     "execution_count": 6,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "def quad(a , b , c , x): return a*x**2 + b*x + c\n",
+    "\n",
+    "quad(3 , 2 , 1 , 1.5)\n",
+    "\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 8,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "10.75"
+      ]
+     },
+     "execution_count": 8,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "from functools import partial\n",
+    "\n",
+    "def mk_quad(a , b , c): return partial(quad ,a , b , c)\n",
+    "\n",
+    "f = mk_quad(3 , 2 , 1)\n",
+    "\n",
+    "f(1.5)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 10,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import torch\n",
+    "import matplotlib.pyplot as plt \n",
+    "def plot_function(f, title=None, min=-2.1, max=2.1, color='r', ylim=None):\n",
+    "    x = torch.linspace(min,max, 100)[:,None]\n",
+    "    if ylim: plt.ylim(ylim)\n",
+    "    plt.plot(x, f(x), color)\n",
+    "    if title is not None: plt.title(title)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 9,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "def mse(preds , acts): return ((preds - acts)**2).mean()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 13,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "def noise(x, scale): return np.random.normal(scale=scale, size=x.shape)\n",
+    "def add_noise(x, mult, add): return x * (1+noise(x,mult)) + noise(x,add)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 14,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import numpy as np\n",
+    "np.random.seed(42)\n",
+    "\n",
+    "x = torch.linspace(-2, 2, steps=20)[:,None]\n",
+    "y = add_noise(f(x), 0.15, 1.5)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 15,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "80a6fa4ee63e4e1c8cf49638c80adf50",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "interactive(children=(FloatSlider(value=1.5, description='a', max=4.5, min=-1.5), FloatSlider(value=1.5, descr…"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    }
+   ],
+   "source": [
+    "from ipywidgets import interact\n",
+    "@interact(a=1.5 , b = 1.5 , c=1.5)\n",
+    "def plot_quad(a , b , c):\n",
+    "    f = mk_quad(a , b , c)\n",
+    "    plt.scatter(x , y)\n",
+    "    loss = mse(f(x) , y)\n",
+    "    plot_function(f , ylim=(-3 , 12) , title = f\"MSE: {loss:.2f}\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 16,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "def quad_mse(params):\n",
+    "    f = mk_quad(*params)\n",
+    "    return mse(f(x) , y)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 20,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "tensor([1.5000, 1.5000, 1.5000], requires_grad=True)"
+      ]
+     },
+     "execution_count": 20,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "abc = torch.tensor([1.5 , 1.5 , 1.5])\n",
+    "abc.requires_grad_()\n",
+    "\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 21,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "tensor(5.4892, dtype=torch.float64, grad_fn=<MeanBackward0>)"
+      ]
+     },
+     "execution_count": 21,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "loss = quad_mse(abc)\n",
+    "\n",
+    "loss"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 22,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "loss.backward()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 23,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "tensor([-7.0908,  1.0602, -1.8620])"
+      ]
+     },
+     "execution_count": 23,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "abc.grad"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 27,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "LOSS IS tensor(2.6496, dtype=torch.float64, grad_fn=<MeanBackward0>)\n",
+      "LOSS IS tensor(3.3536, dtype=torch.float64, grad_fn=<MeanBackward0>)\n",
+      "LOSS IS tensor(4.1631, dtype=torch.float64, grad_fn=<MeanBackward0>)\n",
+      "LOSS IS tensor(4.7701, dtype=torch.float64, grad_fn=<MeanBackward0>)\n",
+      "LOSS IS tensor(4.9408, dtype=torch.float64, grad_fn=<MeanBackward0>)\n",
+      "LOSS IS tensor(4.5979, dtype=torch.float64, grad_fn=<MeanBackward0>)\n",
+      "LOSS IS tensor(3.8490, dtype=torch.float64, grad_fn=<MeanBackward0>)\n",
+      "LOSS IS tensor(2.9481, dtype=torch.float64, grad_fn=<MeanBackward0>)\n",
+      "LOSS IS tensor(2.2058, dtype=torch.float64, grad_fn=<MeanBackward0>)\n",
+      "LOSS IS tensor(1.8796, dtype=torch.float64, grad_fn=<MeanBackward0>)\n",
+      "LOSS IS tensor(2.0819, dtype=torch.float64, grad_fn=<MeanBackward0>)\n",
+      "LOSS IS tensor(2.7403, dtype=torch.float64, grad_fn=<MeanBackward0>)\n",
+      "LOSS IS tensor(3.6224, dtype=torch.float64, grad_fn=<MeanBackward0>)\n",
+      "LOSS IS tensor(4.4176, dtype=torch.float64, grad_fn=<MeanBackward0>)\n",
+      "LOSS IS tensor(4.8469, dtype=torch.float64, grad_fn=<MeanBackward0>)\n",
+      "LOSS IS tensor(4.7612, dtype=torch.float64, grad_fn=<MeanBackward0>)\n",
+      "LOSS IS tensor(4.1943, dtype=torch.float64, grad_fn=<MeanBackward0>)\n",
+      "LOSS IS tensor(3.3515, dtype=torch.float64, grad_fn=<MeanBackward0>)\n",
+      "LOSS IS tensor(2.5374, dtype=torch.float64, grad_fn=<MeanBackward0>)\n",
+      "LOSS IS tensor(2.0492, dtype=torch.float64, grad_fn=<MeanBackward0>)\n",
+      "LOSS IS tensor(2.0717, dtype=torch.float64, grad_fn=<MeanBackward0>)\n",
+      "LOSS IS tensor(2.6126, dtype=torch.float64, grad_fn=<MeanBackward0>)\n",
+      "LOSS IS tensor(3.4991, dtype=torch.float64, grad_fn=<MeanBackward0>)\n",
+      "LOSS IS tensor(4.4391, dtype=torch.float64, grad_fn=<MeanBackward0>)\n",
+      "LOSS IS tensor(5.1229, dtype=torch.float64, grad_fn=<MeanBackward0>)\n",
+      "LOSS IS tensor(5.3318, dtype=torch.float64, grad_fn=<MeanBackward0>)\n",
+      "LOSS IS tensor(5.0142, dtype=torch.float64, grad_fn=<MeanBackward0>)\n",
+      "LOSS IS tensor(4.3023, dtype=torch.float64, grad_fn=<MeanBackward0>)\n",
+      "LOSS IS tensor(3.4640, dtype=torch.float64, grad_fn=<MeanBackward0>)\n",
+      "LOSS IS tensor(2.8078, dtype=torch.float64, grad_fn=<MeanBackward0>)\n",
+      "LOSS IS tensor(2.5724, dtype=torch.float64, grad_fn=<MeanBackward0>)\n",
+      "LOSS IS tensor(2.8419, dtype=torch.float64, grad_fn=<MeanBackward0>)\n",
+      "LOSS IS tensor(3.5167, dtype=torch.float64, grad_fn=<MeanBackward0>)\n",
+      "LOSS IS tensor(4.3478, dtype=torch.float64, grad_fn=<MeanBackward0>)\n",
+      "LOSS IS tensor(5.0262, dtype=torch.float64, grad_fn=<MeanBackward0>)\n",
+      "LOSS IS tensor(5.2927, dtype=torch.float64, grad_fn=<MeanBackward0>)\n",
+      "LOSS IS tensor(5.0303, dtype=torch.float64, grad_fn=<MeanBackward0>)\n",
+      "LOSS IS tensor(4.3074, dtype=torch.float64, grad_fn=<MeanBackward0>)\n",
+      "LOSS IS tensor(3.3546, dtype=torch.float64, grad_fn=<MeanBackward0>)\n",
+      "LOSS IS tensor(2.4847, dtype=torch.float64, grad_fn=<MeanBackward0>)\n",
+      "LOSS IS tensor(1.9837, dtype=torch.float64, grad_fn=<MeanBackward0>)\n",
+      "LOSS IS tensor(2.0101, dtype=torch.float64, grad_fn=<MeanBackward0>)\n",
+      "LOSS IS tensor(2.5399, dtype=torch.float64, grad_fn=<MeanBackward0>)\n",
+      "LOSS IS tensor(3.3748, dtype=torch.float64, grad_fn=<MeanBackward0>)\n",
+      "LOSS IS tensor(4.2127, dtype=torch.float64, grad_fn=<MeanBackward0>)\n",
+      "LOSS IS tensor(4.7532, dtype=torch.float64, grad_fn=<MeanBackward0>)\n",
+      "LOSS IS tensor(4.8036, dtype=torch.float64, grad_fn=<MeanBackward0>)\n",
+      "LOSS IS tensor(4.3463, dtype=torch.float64, grad_fn=<MeanBackward0>)\n",
+      "LOSS IS tensor(3.5442, dtype=torch.float64, grad_fn=<MeanBackward0>)\n",
+      "LOSS IS tensor(2.6829, dtype=torch.float64, grad_fn=<MeanBackward0>)\n"
+     ]
+    }
+   ],
+   "source": [
+    "for i in range(50):\n",
+    "    loss = quad_mse(abc)\n",
+    "    loss.backward()\n",
+    "    with torch.no_grad():\n",
+    "        abc -= abc.grad * 0.01\n",
+    "    print(\"LOSS IS\" , loss)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 28,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "tensor([2.8198, 0.7940, 0.9271], requires_grad=True)"
+      ]
+     },
+     "execution_count": 28,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "abc"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 29,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "def rectified_linear(m , b , x):\n",
+    "    y = m*x + b\n",
+    "    return torch.clip(y , 0.0)\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "diffusers",
+   "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": "3.8.16"
+  },
+  "orig_nbformat": 4,
+  "vscode": {
+   "interpreter": {
+    "hash": "d26bfdf1b33eed3ee2dd554c3e2dd92c45e09514ce35d65c074828f753b40123"
+   }
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 2
+}