09b_learner_jcat.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Let's kill off `Runner`"
   ]
  },
  {
   "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": [],
   "source": [
    "#export\n",
    "from exp.nb_09 import *"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "exp.nb_04.AvgStats"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "AvgStats"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Imagenette data"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "[Jump_to lesson 11 video](https://course.fast.ai/videos/?lesson=11&t=6571)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "path = datasets.untar_data(datasets.URLs.IMAGENETTE_160)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "tfms = [make_rgb, ResizeFixed(128), to_byte_tensor, to_float_tensor]\n",
    "bs=64\n",
    "\n",
    "il = ImageList.from_files(path, tfms=tfms)\n",
    "sd = SplitData.split_by_func(il, partial(grandparent_splitter, valid_name='val'))\n",
    "ll = label_by_func(sd, parent_labeler, proc_y=CategoryProcessor())\n",
    "data = ll.to_databunch(bs, c_in=3, c_out=10, num_workers=4)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "cbfs = [partial(AvgStatsCallback,accuracy),\n",
    "        CudaCallback,\n",
    "        partial(BatchTransformXCallback, norm_imagenette)]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "nfs = [32]*4"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Having a Runner is great but not essential when the `Learner` already has everything needed in its state. We implement everything inside it directly instead of building a second object."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### In Lesson 12 Jeremy Howard revisited material in the cell below  [Jump_to lesson 12 video](https://course.fast.ai/videos/?lesson=12&t=65)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "#export\n",
    "def param_getter(m): return m.parameters()\n",
    "\n",
    "class Learner():\n",
    "    def __init__(self, model, data, loss_func, opt_func=sgd_opt, lr=1e-2, splitter=param_getter,\n",
    "                 cbs=None, cb_funcs=None):\n",
    "        self.model,self.data,self.loss_func,self.opt_func,self.lr,self.splitter = model,data,loss_func,opt_func,lr,splitter\n",
    "        self.in_train,self.logger,self.opt = False,print,None\n",
    "        \n",
    "        # NB: Things marked \"NEW\" are covered in lesson 12\n",
    "        # NEW: avoid need for set_runner\n",
    "        self.cbs = []\n",
    "        self.add_cb(TrainEvalCallback())\n",
    "        self.add_cbs(cbs)\n",
    "        self.add_cbs(cbf() for cbf in listify(cb_funcs))\n",
    "\n",
    "    def add_cbs(self, cbs):\n",
    "        for cb in listify(cbs): self.add_cb(cb)\n",
    "            \n",
    "    def add_cb(self, cb):\n",
    "        cb.set_runner(self)\n",
    "        setattr(self, cb.name, cb)\n",
    "        self.cbs.append(cb)\n",
    "\n",
    "    def remove_cbs(self, cbs):\n",
    "        for cb in listify(cbs): self.cbs.remove(cb)\n",
    "            \n",
    "    def one_batch(self, i, xb, yb):\n",
    "        try:\n",
    "            self.iter = i\n",
    "            self.xb,self.yb = xb,yb;                        self('begin_batch')\n",
    "            self.pred = self.model(self.xb);                self('after_pred')\n",
    "            self.loss = self.loss_func(self.pred, self.yb); self('after_loss')\n",
    "            if not self.in_train: return\n",
    "            self.loss.backward();                           self('after_backward')\n",
    "            self.opt.step();                                self('after_step')\n",
    "            self.opt.zero_grad()\n",
    "        except CancelBatchException:                        self('after_cancel_batch')\n",
    "        finally:                                            self('after_batch')\n",
    "\n",
    "    def all_batches(self):\n",
    "        self.iters = len(self.dl)\n",
    "        try:\n",
    "            for i,(xb,yb) in enumerate(self.dl): self.one_batch(i, xb, yb)\n",
    "        except CancelEpochException: self('after_cancel_epoch')\n",
    "\n",
    "    def do_begin_fit(self, epochs):\n",
    "        self.epochs,self.loss = epochs,tensor(0.)\n",
    "        self('begin_fit')\n",
    "\n",
    "    def do_begin_epoch(self, epoch):\n",
    "        self.epoch,self.dl = epoch,self.data.train_dl\n",
    "        return self('begin_epoch')\n",
    "\n",
    "    def fit(self, epochs, cbs=None, reset_opt=False):\n",
    "        # NEW: pass callbacks to fit() and have them removed when done\n",
    "        self.add_cbs(cbs)\n",
    "        # NEW: create optimizer on fit(), optionally replacing existing\n",
    "        if reset_opt or not self.opt: self.opt = self.opt_func(self.splitter(self.model), lr=self.lr)\n",
    "            \n",
    "        try:\n",
    "            self.do_begin_fit(epochs)\n",
    "            for epoch in range(epochs):\n",
    "                self.do_begin_epoch(epoch)\n",
    "                if not self('begin_epoch'): self.all_batches()\n",
    "\n",
    "                with torch.no_grad(): \n",
    "                    self.dl = self.data.valid_dl\n",
    "                    if not self('begin_validate'): self.all_batches()\n",
    "                self('after_epoch')\n",
    "            \n",
    "        except CancelTrainException: self('after_cancel_train')\n",
    "        finally:\n",
    "            self('after_fit')\n",
    "            self.remove_cbs(cbs)\n",
    "\n",
    "    ALL_CBS = {'begin_batch', 'after_pred', 'after_loss', 'after_backward', 'after_step',\n",
    "        'after_cancel_batch', 'after_batch', 'after_cancel_epoch', 'begin_fit',\n",
    "        'begin_epoch', 'begin_validate', 'after_epoch',\n",
    "        'after_cancel_train', 'after_fit'}\n",
    "    \n",
    "    def __call__(self, cb_name):\n",
    "        res = False\n",
    "        assert cb_name in self.ALL_CBS\n",
    "        for cb in sorted(self.cbs, key=lambda x: x._order): res = cb(cb_name) and res\n",
    "        return res"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "#export\n",
    "class AvgStatsCallback(Callback):\n",
    "    def __init__(self, metrics):\n",
    "        self.train_stats,self.valid_stats = AvgStats(metrics,True),AvgStats(metrics,False)\n",
    "        \n",
    "    def begin_epoch(self):\n",
    "        self.train_stats.reset()\n",
    "        self.valid_stats.reset()\n",
    "        \n",
    "    def after_loss(self):\n",
    "        stats = self.train_stats if self.in_train else self.valid_stats\n",
    "        with torch.no_grad(): stats.accumulate(self.run)\n",
    "    \n",
    "    def after_epoch(self):\n",
    "        #We use the logger function of the `Learner` here, it can be customized to write in a file or in a progress bar\n",
    "        self.logger(self.train_stats)\n",
    "        self.logger(self.valid_stats) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "cbfs = [partial(AvgStatsCallback,accuracy),\n",
    "        CudaCallback,\n",
    "        partial(BatchTransformXCallback, norm_imagenette)]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "#export\n",
    "def get_learner(nfs, data, lr, layer, loss_func=F.cross_entropy,\n",
    "                cb_funcs=None, opt_func=sgd_opt, **kwargs):\n",
    "    model = get_cnn_model(data, nfs, layer, **kwargs)\n",
    "    init_cnn(model)\n",
    "    return Learner(model, data, loss_func, lr=lr, cb_funcs=cb_funcs, opt_func=opt_func)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "learn = get_learner(nfs, data, 0.4, conv_layer, cb_funcs=cbfs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train: [1.733643221895843, tensor(0.4024, device='cuda:0')]\n",
      "valid: [1.5155780029296875, tensor(0.4860, device='cuda:0')]\n",
      "Wall time: 45.2 s\n"
     ]
    }
   ],
   "source": [
    "%time learn.fit(1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Check everything works"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's check our previous callbacks still work."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "cbfs += [Recorder]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "learn = get_learner(nfs, data, 0.4, conv_layer, cb_funcs=cbfs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "phases = combine_scheds([0.3, 0.7], cos_1cycle_anneal(0.2, 0.6, 0.2))\n",
    "sched = ParamScheduler('lr', phases)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train: [1.751972513863037, tensor(0.3890, device='cuda:0')]\n",
      "valid: [1.4409005126953125, tensor(0.4960, device='cuda:0')]\n"
     ]
    }
   ],
   "source": [
    "learn.fit(1, sched)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "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": [
    "learn.recorder.plot_lr()"
   ]
  },
  {
   "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": [
    "learn.recorder.plot_loss()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Export"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Converted 09b_learner.ipynb to exp\\nb_09b.py\n"
     ]
    }
   ],
   "source": [
    "# !./notebook2script.py 09b_learner.ipynb\n",
    "!python notebook2script.py 09b_learner.ipynb"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "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": "3.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}