_05_anneal (Skipped).ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "%load_ext autoreload\n",
    "%autoreload 2\n",
    "\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "importing Jupyter notebook from notebook2script_py.ipynb\n"
     ]
    }
   ],
   "source": [
    "#export\n",
    "from exp.nb_04 import *"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Initial setup"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "[Jump_to lesson 9 video](https://course19.fast.ai/videos/?lesson=9&t=7013)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "x_train,y_train,x_valid,y_valid = get_data()\n",
    "y_train = y_train.type(torch.LongTensor)\n",
    "y_valid = y_valid.type(torch.LongTensor)\n",
    "train_ds,valid_ds = Dataset(x_train, y_train),Dataset(x_valid, y_valid)\n",
    "nh,bs = 50,512\n",
    "c = y_train.max().item()+1\n",
    "loss_func = F.cross_entropy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "data = DataBunch(*get_dls(train_ds, valid_ds, bs), c)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "#export\n",
    "def create_learner(model_func, loss_func, data):\n",
    "    return Learner(*model_func(data), loss_func, data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train: [0.6780896875, tensor(0.8007)]\n",
      "valid: [0.312122509765625, tensor(0.9111)]\n",
      "train: [0.30380400390625, tensor(0.9119)]\n",
      "valid: [0.261823779296875, tensor(0.9231)]\n",
      "train: [0.2486877734375, tensor(0.9280)]\n",
      "valid: [0.2070280517578125, tensor(0.9426)]\n"
     ]
    }
   ],
   "source": [
    "learn = create_learner(get_model, loss_func, data)\n",
    "run = Runner([AvgStatsCallback([accuracy])])\n",
    "\n",
    "run.fit(3, learn)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train: [0.75576359375, tensor(0.8058)]\n",
      "valid: [0.3507890625, tensor(0.9046)]\n",
      "train: [0.3428965625, tensor(0.9022)]\n",
      "valid: [0.295852685546875, tensor(0.9148)]\n",
      "train: [0.2905862109375, tensor(0.9177)]\n",
      "valid: [0.251711376953125, tensor(0.9295)]\n"
     ]
    }
   ],
   "source": [
    "learn = create_learner(partial(get_model, lr=0.3), loss_func, data)\n",
    "run = Runner([AvgStatsCallback([accuracy])])\n",
    "\n",
    "run.fit(3, learn)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "#export\n",
    "def get_model_func(lr=0.5): return partial(get_model, lr=lr)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Annealing"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We define two new callbacks: the Recorder to save track of the loss and our scheduled learning rate, and a ParamScheduler that can schedule any hyperparameter as long as it's registered in the state_dict of the optimizer. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "[Jump_to lesson 9 video](https://course19.fast.ai/videos/?lesson=9&t=7202)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "#export\n",
    "class Recorder(Callback):\n",
    "    def begin_fit(self): self.lrs,self.losses = [],[]\n",
    "\n",
    "    def after_batch(self):\n",
    "        if not self.in_train: return\n",
    "        self.lrs.append(self.opt.param_groups[-1]['lr'])\n",
    "        self.losses.append(self.loss.detach().cpu())        \n",
    "\n",
    "    def plot_lr  (self): plt.plot(self.lrs)\n",
    "    def plot_loss(self): plt.plot(self.losses)\n",
    "\n",
    "class ParamScheduler(Callback):\n",
    "    _order=1\n",
    "    def __init__(self, pname, sched_func): self.pname,self.sched_func = pname,sched_func\n",
    "\n",
    "    def set_param(self):\n",
    "        for pg in self.opt.param_groups:\n",
    "            pg[self.pname] = self.sched_func(self.n_epochs/self.epochs)\n",
    "            \n",
    "    def begin_batch(self): \n",
    "        if self.in_train: self.set_param()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's start with a simple linear schedule going from start to end. It returns a function that takes a `pos` argument (going from 0 to 1) such that this function goes from `start` (at `pos=0`) to `end` (at `pos=1`) in a linear fashion."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "[Jump_to lesson 9 video](https://course19.fast.ai/videos/?lesson=9&t=7431)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "def sched_lin(start, end):\n",
    "    def _inner(start, end, pos): return start + pos*(end-start)\n",
    "    return partial(_inner, start, end)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We can refactor this with a decorator."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "[Jump_to lesson 9 video](https://course19.fast.ai/videos/?lesson=9&t=7526)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "#export\n",
    "def annealer(f):\n",
    "    def _inner(start, end): return partial(f, start, end)\n",
    "    return _inner\n",
    "\n",
    "@annealer\n",
    "def sched_lin(start, end, pos): return start + pos*(end-start)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "# shift-tab works too, in Jupyter!\n",
    "#sched_lin()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1.3"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "f = sched_lin(1,2)\n",
    "f(0.3)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "And here are other scheduler functions:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "#export\n",
    "@annealer\n",
    "def sched_cos(start, end, pos): return start + (1 + math.cos(math.pi*(1-pos))) * (end-start) / 2\n",
    "@annealer\n",
    "def sched_no(start, end, pos):  return start\n",
    "@annealer\n",
    "def sched_exp(start, end, pos): return start * (end/start) ** pos\n",
    "\n",
    "def cos_1cycle_anneal(start, high, end):\n",
    "    return [sched_cos(start, high), sched_cos(high, end)]\n",
    "\n",
    "#This monkey-patch is there to be able to plot tensors\n",
    "torch.Tensor.ndim = property(lambda x: len(x.shape))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "[Jump_to lesson 9 video](https://course19.fast.ai/videos/?lesson=9&t=7730)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "annealings = \"NO LINEAR COS EXP\".split()\n",
    "\n",
    "a = torch.arange(0, 100)\n",
    "p = torch.linspace(0.01,1,100)\n",
    "\n",
    "fns = [sched_no, sched_lin, sched_cos, sched_exp]\n",
    "for fn, t in zip(fns, annealings):\n",
    "    f = fn(2, 1e-2)\n",
    "    plt.plot(a, [f(o) for o in p], label=t)\n",
    "plt.legend();"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In practice, we'll often want to combine different schedulers, the following function does that: it uses `scheds[i]` for `pcts[i]` of the training."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "#export\n",
    "def combine_scheds(pcts, scheds):\n",
    "    assert sum(pcts) == 1.\n",
    "    pcts = tensor([0] + listify(pcts))\n",
    "    assert torch.all(pcts >= 0)\n",
    "    pcts = torch.cumsum(pcts, 0)\n",
    "    def _inner(pos):\n",
    "        idx = (pos >= pcts).nonzero().max()\n",
    "        if idx == 2: idx = 1\n",
    "        actual_pos = (pos-pcts[idx]) / (pcts[idx+1]-pcts[idx])\n",
    "        return scheds[idx](actual_pos)\n",
    "    return _inner"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Here is an example: use 30% of the budget to go from 0.3 to 0.6 following a cosine, then the last 70% of the budget to go from 0.6 to 0.2, still following a cosine."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "sched = combine_scheds([0.3, 0.7], [sched_cos(0.3, 0.6), sched_cos(0.6, 0.2)]) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x1bc9b3926c8>]"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(a, [sched(o) for o in p])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We can use it for training quite easily..."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "cbfs = [Recorder,\n",
    "        partial(AvgStatsCallback,accuracy),\n",
    "        partial(ParamScheduler, 'lr', sched)]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "learn = create_learner(get_model_func(0.3), loss_func, data)\n",
    "run = Runner(cb_funcs=cbfs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train: [0.781123515625, tensor(0.7879)]\n",
      "valid: [0.349321435546875, tensor(0.8949)]\n",
      "train: [0.2943194921875, tensor(0.9135)]\n",
      "valid: [0.232355126953125, tensor(0.9347)]\n",
      "train: [0.23255744140625, tensor(0.9335)]\n",
      "valid: [0.20632099609375, tensor(0.9428)]\n"
     ]
    }
   ],
   "source": [
    "run.fit(3, learn)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "... then check with our recorder if the learning rate followed the right schedule."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "run.recorder.plot_lr()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "run.recorder.plot_loss()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Reference for this notebook"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#https://github.com/fastai/course-v3/blob/master/nbs/dl2/05_anneal.ipynb"
   ]
  }
 ],
 "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": "3.7.11"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}