gradient_noise.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "6f1f5335-17a1-4c87-af5b-e36d67cf7b72",
   "metadata": {},
   "outputs": [],
   "source": [
    "%load_ext autoreload\n",
    "%autoreload 2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "3ffa30c8-6928-4ed5-9e32-3157fcd498bb",
   "metadata": {},
   "outputs": [],
   "source": [
    "%run imports.py"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "73a05f30-3cef-41f9-bd88-e15de4faa748",
   "metadata": {},
   "outputs": [],
   "source": [
    "import util"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "76b13b43-79e8-4648-aeb5-8d30610c558c",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Files already downloaded and verified\n",
      "Files already downloaded and verified\n"
     ]
    }
   ],
   "source": [
    "ds = cifar10.CIFAR10()\n",
    "# ds = mnist.MNIST()\n",
    "model = cifar10.BigNetwork"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "418b48af-b61b-43ac-814e-08ed22f38801",
   "metadata": {},
   "outputs": [],
   "source": [
    "device = torch.device(\"cuda\") if torch.cuda.is_available() else torch.device(\"cpu\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "3d007b36-ca35-4988-9977-405f99d71ee4",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Files already downloaded and verified\n",
      "Files already downloaded and verified\n"
     ]
    }
   ],
   "source": [
    "\n",
    "ds_small = cifar10.CIFAR10(bs_train=10, bs_test=100)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "id": "38092a76-f357-4b5e-bdda-c2e3067a46aa",
   "metadata": {},
   "outputs": [],
   "source": [
    "def calc_grad_noise(net):\n",
    "    opt = torch.optim.Adam(net.parameters(), lr=0.)\n",
    "\n",
    "    gs = []\n",
    "    for i in range(50):\n",
    "        x, y = next(iter(ds_small.dl_train))\n",
    "        yp = net(x)\n",
    "        loss = ds.loss_fn(yp, y)\n",
    "\n",
    "        opt.zero_grad()\n",
    "        loss.backward()\n",
    "        \n",
    "        g = torch.cat([pi.grad.flatten() for pi in net.parameters()])\n",
    "        gs.append(g)\n",
    "    gs = torch.stack(gs)\n",
    "    dots = (gs[:, None]*gs[None, :]).sum(dim=-1)\n",
    "    \n",
    "    return (dots*(1-torch.eye(len(dots)))).mean().item()\n",
    "    \n",
    "    \n",
    "def train_adam(net, accu, lr=1e-2):\n",
    "    opt = torch.optim.Adam(net.parameters(), lr=lr)\n",
    "    \n",
    "    for epoch_idx in tqdm(range(5)):\n",
    "        loop = tqdm(ds.dl_train)\n",
    "        for x, y in loop:\n",
    "            grad_noise = calc_grad_noise(net)\n",
    "            # print(grad_noise)\n",
    "            \n",
    "            yp = net(x)\n",
    "            loss = ds.loss_fn(yp, y)\n",
    "            \n",
    "            opt.zero_grad()\n",
    "            loss.backward()\n",
    "            opt.step()\n",
    "            data = dict(loss_batch=loss.item(), grad_noise=grad_noise)\n",
    "            accu.append(**data)\n",
    "            loop.set_postfix(data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "id": "7cd306f3-1de1-47b1-a13b-ab32f419ba8d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "44d2461c88274bed8245d49b9d2227ef",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/5 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "0fe8660be02142f1a7dfc2d7609b74ca",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/50 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "e94c1ee15caf40ae9b98bacc585ca8cc",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/50 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "accef21b8b1f41a788fa10b82d67c761",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/50 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "4d5997e0b5634549b9afa9ddc143d07f",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/50 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "7d3e736aaa3f4754b31bc4cd9fc52eeb",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/50 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "net = model()\n",
    "accu = util.Accumulator()\n",
    "train_adam(net, accu)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "eb3e06e9-294d-41b9-9d06-d8b1bee55f29",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "id": "cfbebb74-0721-416b-91c9-9e4c18ce255d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "util.plot_accumulators([accu])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "id": "7270f02a-6c5a-439f-98ea-6e9c639def81",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<util.Accumulator at 0x2abf4de40430>"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "accu"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "72f24320-5247-4b0a-988d-b320819fff35",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "f2805cfd-d085-4187-89df-b7bdc7b60259",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "a0779875-ae1f-42fe-837f-bd2526aa5067",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "bc01bd80-896f-47dc-bb6b-88750d96df23",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "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": "3.9.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}