Jeremy Janella
153 lines
16 KiB
Plaintext
153 lines
16 KiB
Plaintext
{
|
|
"cells": [
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 1,
|
|
"id": "5373faae-c4aa-4f05-bbd7-4b7728215aea",
|
|
"metadata": {},
|
|
"outputs": [
|
|
{
|
|
"ename": "RuntimeError",
|
|
"evalue": "operator torchvision::nms does not exist",
|
|
"output_type": "error",
|
|
"traceback": [
|
|
"\u001b[31m---------------------------------------------------------------------------\u001b[39m",
|
|
"\u001b[31mRuntimeError\u001b[39m Traceback (most recent call last)",
|
|
"\u001b[36mCell\u001b[39m\u001b[36m \u001b[39m\u001b[32mIn[1]\u001b[39m\u001b[32m, line 5\u001b[39m\n\u001b[32m 3\u001b[39m \u001b[38;5;28;01mimport\u001b[39;00m\u001b[38;5;250m \u001b[39m\u001b[34;01mtorch\u001b[39;00m\u001b[34;01m.\u001b[39;00m\u001b[34;01mnn\u001b[39;00m\u001b[34;01m.\u001b[39;00m\u001b[34;01mfunctional\u001b[39;00m\u001b[38;5;250m \u001b[39m\u001b[38;5;28;01mas\u001b[39;00m\u001b[38;5;250m \u001b[39m\u001b[34;01mF\u001b[39;00m\n\u001b[32m 4\u001b[39m \u001b[38;5;28;01mfrom\u001b[39;00m\u001b[38;5;250m \u001b[39m\u001b[34;01mtorch\u001b[39;00m\u001b[34;01m.\u001b[39;00m\u001b[34;01mutils\u001b[39;00m\u001b[34;01m.\u001b[39;00m\u001b[34;01mdata\u001b[39;00m\u001b[38;5;250m \u001b[39m\u001b[38;5;28;01mimport\u001b[39;00m DataLoader\n\u001b[32m----> \u001b[39m\u001b[32m5\u001b[39m \u001b[38;5;28;01mfrom\u001b[39;00m\u001b[38;5;250m \u001b[39m\u001b[34;01mtorchvision\u001b[39;00m\u001b[38;5;250m \u001b[39m\u001b[38;5;28;01mimport\u001b[39;00m datasets, transforms\n\u001b[32m 6\u001b[39m \u001b[38;5;28;01mimport\u001b[39;00m\u001b[38;5;250m \u001b[39m\u001b[34;01mnumpy\u001b[39;00m\u001b[38;5;250m \u001b[39m\u001b[38;5;28;01mas\u001b[39;00m\u001b[38;5;250m \u001b[39m\u001b[34;01mnp\u001b[39;00m\n\u001b[32m 7\u001b[39m \u001b[38;5;28;01mimport\u001b[39;00m\u001b[38;5;250m \u001b[39m\u001b[34;01mmatplotlib\u001b[39;00m\u001b[34;01m.\u001b[39;00m\u001b[34;01mpyplot\u001b[39;00m\u001b[38;5;250m \u001b[39m\u001b[38;5;28;01mas\u001b[39;00m\u001b[38;5;250m \u001b[39m\u001b[34;01mplt\u001b[39;00m\n",
|
|
"\u001b[36mFile \u001b[39m\u001b[32m~/.conda/envs/ai/lib/python3.11/site-packages/torchvision/__init__.py:10\u001b[39m\n\u001b[32m 7\u001b[39m \u001b[38;5;66;03m# Don't re-order these, we need to load the _C extension (done when importing\u001b[39;00m\n\u001b[32m 8\u001b[39m \u001b[38;5;66;03m# .extensions) before entering _meta_registrations.\u001b[39;00m\n\u001b[32m 9\u001b[39m \u001b[38;5;28;01mfrom\u001b[39;00m\u001b[38;5;250m \u001b[39m\u001b[34;01m.\u001b[39;00m\u001b[34;01mextension\u001b[39;00m\u001b[38;5;250m \u001b[39m\u001b[38;5;28;01mimport\u001b[39;00m _HAS_OPS \u001b[38;5;66;03m# usort:skip\u001b[39;00m\n\u001b[32m---> \u001b[39m\u001b[32m10\u001b[39m \u001b[38;5;28;01mfrom\u001b[39;00m\u001b[38;5;250m \u001b[39m\u001b[34;01mtorchvision\u001b[39;00m\u001b[38;5;250m \u001b[39m\u001b[38;5;28;01mimport\u001b[39;00m _meta_registrations, datasets, io, models, ops, transforms, utils \u001b[38;5;66;03m# usort:skip\u001b[39;00m\n\u001b[32m 12\u001b[39m \u001b[38;5;28;01mtry\u001b[39;00m:\n\u001b[32m 13\u001b[39m \u001b[38;5;28;01mfrom\u001b[39;00m\u001b[38;5;250m \u001b[39m\u001b[34;01m.\u001b[39;00m\u001b[34;01mversion\u001b[39;00m\u001b[38;5;250m \u001b[39m\u001b[38;5;28;01mimport\u001b[39;00m __version__ \u001b[38;5;66;03m# noqa: F401\u001b[39;00m\n",
|
|
"\u001b[36mFile \u001b[39m\u001b[32m~/.conda/envs/ai/lib/python3.11/site-packages/torchvision/_meta_registrations.py:163\u001b[39m\n\u001b[32m 153\u001b[39m torch._check(\n\u001b[32m 154\u001b[39m grad.dtype == rois.dtype,\n\u001b[32m 155\u001b[39m \u001b[38;5;28;01mlambda\u001b[39;00m: (\n\u001b[32m (...)\u001b[39m\u001b[32m 158\u001b[39m ),\n\u001b[32m 159\u001b[39m )\n\u001b[32m 160\u001b[39m \u001b[38;5;28;01mreturn\u001b[39;00m grad.new_empty((batch_size, channels, height, width))\n\u001b[32m--> \u001b[39m\u001b[32m163\u001b[39m \u001b[38;5;129;43m@torch\u001b[39;49m\u001b[43m.\u001b[49m\u001b[43mlibrary\u001b[49m\u001b[43m.\u001b[49m\u001b[43mregister_fake\u001b[49m\u001b[43m(\u001b[49m\u001b[33;43m\"\u001b[39;49m\u001b[33;43mtorchvision::nms\u001b[39;49m\u001b[33;43m\"\u001b[39;49m\u001b[43m)\u001b[49m\n\u001b[32m 164\u001b[39m \u001b[38;5;28;43;01mdef\u001b[39;49;00m\u001b[38;5;250;43m \u001b[39;49m\u001b[34;43mmeta_nms\u001b[39;49m\u001b[43m(\u001b[49m\u001b[43mdets\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mscores\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43miou_threshold\u001b[49m\u001b[43m)\u001b[49m\u001b[43m:\u001b[49m\n\u001b[32m 165\u001b[39m \u001b[43m \u001b[49m\u001b[43mtorch\u001b[49m\u001b[43m.\u001b[49m\u001b[43m_check\u001b[49m\u001b[43m(\u001b[49m\u001b[43mdets\u001b[49m\u001b[43m.\u001b[49m\u001b[43mdim\u001b[49m\u001b[43m(\u001b[49m\u001b[43m)\u001b[49m\u001b[43m \u001b[49m\u001b[43m==\u001b[49m\u001b[43m \u001b[49m\u001b[32;43m2\u001b[39;49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[38;5;28;43;01mlambda\u001b[39;49;00m\u001b[43m:\u001b[49m\u001b[43m \u001b[49m\u001b[33;43mf\u001b[39;49m\u001b[33;43m\"\u001b[39;49m\u001b[33;43mboxes should be a 2d tensor, got \u001b[39;49m\u001b[38;5;132;43;01m{\u001b[39;49;00m\u001b[43mdets\u001b[49m\u001b[43m.\u001b[49m\u001b[43mdim\u001b[49m\u001b[43m(\u001b[49m\u001b[43m)\u001b[49m\u001b[38;5;132;43;01m}\u001b[39;49;00m\u001b[33;43mD\u001b[39;49m\u001b[33;43m\"\u001b[39;49m\u001b[43m)\u001b[49m\n\u001b[32m 166\u001b[39m \u001b[43m \u001b[49m\u001b[43mtorch\u001b[49m\u001b[43m.\u001b[49m\u001b[43m_check\u001b[49m\u001b[43m(\u001b[49m\u001b[43mdets\u001b[49m\u001b[43m.\u001b[49m\u001b[43msize\u001b[49m\u001b[43m(\u001b[49m\u001b[32;43m1\u001b[39;49m\u001b[43m)\u001b[49m\u001b[43m \u001b[49m\u001b[43m==\u001b[49m\u001b[43m \u001b[49m\u001b[32;43m4\u001b[39;49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[38;5;28;43;01mlambda\u001b[39;49;00m\u001b[43m:\u001b[49m\u001b[43m \u001b[49m\u001b[33;43mf\u001b[39;49m\u001b[33;43m\"\u001b[39;49m\u001b[33;43mboxes should have 4 elements in dimension 1, got \u001b[39;49m\u001b[38;5;132;43;01m{\u001b[39;49;00m\u001b[43mdets\u001b[49m\u001b[43m.\u001b[49m\u001b[43msize\u001b[49m\u001b[43m(\u001b[49m\u001b[32;43m1\u001b[39;49m\u001b[43m)\u001b[49m\u001b[38;5;132;43;01m}\u001b[39;49;00m\u001b[33;43m\"\u001b[39;49m\u001b[43m)\u001b[49m\n",
|
|
"\u001b[36mFile \u001b[39m\u001b[32m~/.conda/envs/ai/lib/python3.11/site-packages/torch/library.py:1073\u001b[39m, in \u001b[36mregister_fake.<locals>.register\u001b[39m\u001b[34m(func)\u001b[39m\n\u001b[32m 1071\u001b[39m \u001b[38;5;28;01melse\u001b[39;00m:\n\u001b[32m 1072\u001b[39m use_lib = lib\n\u001b[32m-> \u001b[39m\u001b[32m1073\u001b[39m \u001b[43muse_lib\u001b[49m\u001b[43m.\u001b[49m\u001b[43m_register_fake\u001b[49m\u001b[43m(\u001b[49m\n\u001b[32m 1074\u001b[39m \u001b[43m \u001b[49m\u001b[43mop_name\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mfunc\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43m_stacklevel\u001b[49m\u001b[43m=\u001b[49m\u001b[43mstacklevel\u001b[49m\u001b[43m \u001b[49m\u001b[43m+\u001b[49m\u001b[43m \u001b[49m\u001b[32;43m1\u001b[39;49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mallow_override\u001b[49m\u001b[43m=\u001b[49m\u001b[43mallow_override\u001b[49m\n\u001b[32m 1075\u001b[39m \u001b[43m\u001b[49m\u001b[43m)\u001b[49m\n\u001b[32m 1076\u001b[39m \u001b[38;5;28;01mreturn\u001b[39;00m func\n",
|
|
"\u001b[36mFile \u001b[39m\u001b[32m~/.conda/envs/ai/lib/python3.11/site-packages/torch/library.py:203\u001b[39m, in \u001b[36mLibrary._register_fake\u001b[39m\u001b[34m(self, op_name, fn, _stacklevel, allow_override)\u001b[39m\n\u001b[32m 200\u001b[39m \u001b[38;5;28;01melse\u001b[39;00m:\n\u001b[32m 201\u001b[39m func_to_register = fn\n\u001b[32m--> \u001b[39m\u001b[32m203\u001b[39m handle = \u001b[43mentry\u001b[49m\u001b[43m.\u001b[49m\u001b[43mfake_impl\u001b[49m\u001b[43m.\u001b[49m\u001b[43mregister\u001b[49m\u001b[43m(\u001b[49m\n\u001b[32m 204\u001b[39m \u001b[43m \u001b[49m\u001b[43mfunc_to_register\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43msource\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mlib\u001b[49m\u001b[43m=\u001b[49m\u001b[38;5;28;43mself\u001b[39;49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mallow_override\u001b[49m\u001b[43m=\u001b[49m\u001b[43mallow_override\u001b[49m\n\u001b[32m 205\u001b[39m \u001b[43m\u001b[49m\u001b[43m)\u001b[49m\n\u001b[32m 206\u001b[39m \u001b[38;5;28mself\u001b[39m._registration_handles.append(handle)\n",
|
|
"\u001b[36mFile \u001b[39m\u001b[32m~/.conda/envs/ai/lib/python3.11/site-packages/torch/_library/fake_impl.py:50\u001b[39m, in \u001b[36mFakeImplHolder.register\u001b[39m\u001b[34m(self, func, source, lib, allow_override)\u001b[39m\n\u001b[32m 44\u001b[39m \u001b[38;5;28;01mif\u001b[39;00m \u001b[38;5;28mself\u001b[39m.kernel \u001b[38;5;129;01mis\u001b[39;00m \u001b[38;5;129;01mnot\u001b[39;00m \u001b[38;5;28;01mNone\u001b[39;00m:\n\u001b[32m 45\u001b[39m \u001b[38;5;28;01mraise\u001b[39;00m \u001b[38;5;167;01mRuntimeError\u001b[39;00m(\n\u001b[32m 46\u001b[39m \u001b[33mf\u001b[39m\u001b[33m\"\u001b[39m\u001b[33mregister_fake(...): the operator \u001b[39m\u001b[38;5;132;01m{\u001b[39;00m\u001b[38;5;28mself\u001b[39m.qualname\u001b[38;5;132;01m}\u001b[39;00m\u001b[33m \u001b[39m\u001b[33m\"\u001b[39m\n\u001b[32m 47\u001b[39m \u001b[33mf\u001b[39m\u001b[33m\"\u001b[39m\u001b[33malready has an fake impl registered at \u001b[39m\u001b[33m\"\u001b[39m\n\u001b[32m 48\u001b[39m \u001b[33mf\u001b[39m\u001b[33m\"\u001b[39m\u001b[38;5;132;01m{\u001b[39;00m\u001b[38;5;28mself\u001b[39m.kernel.source\u001b[38;5;132;01m}\u001b[39;00m\u001b[33m.\u001b[39m\u001b[33m\"\u001b[39m\n\u001b[32m 49\u001b[39m )\n\u001b[32m---> \u001b[39m\u001b[32m50\u001b[39m \u001b[38;5;28;01mif\u001b[39;00m \u001b[43mtorch\u001b[49m\u001b[43m.\u001b[49m\u001b[43m_C\u001b[49m\u001b[43m.\u001b[49m\u001b[43m_dispatch_has_kernel_for_dispatch_key\u001b[49m\u001b[43m(\u001b[49m\u001b[38;5;28;43mself\u001b[39;49m\u001b[43m.\u001b[49m\u001b[43mqualname\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[33;43m\"\u001b[39;49m\u001b[33;43mMeta\u001b[39;49m\u001b[33;43m\"\u001b[39;49m\u001b[43m)\u001b[49m:\n\u001b[32m 51\u001b[39m \u001b[38;5;28;01mraise\u001b[39;00m \u001b[38;5;167;01mRuntimeError\u001b[39;00m(\n\u001b[32m 52\u001b[39m \u001b[33mf\u001b[39m\u001b[33m\"\u001b[39m\u001b[33mregister_fake(...): the operator \u001b[39m\u001b[38;5;132;01m{\u001b[39;00m\u001b[38;5;28mself\u001b[39m.qualname\u001b[38;5;132;01m}\u001b[39;00m\u001b[33m \u001b[39m\u001b[33m\"\u001b[39m\n\u001b[32m 53\u001b[39m \u001b[33mf\u001b[39m\u001b[33m\"\u001b[39m\u001b[33malready has an DispatchKey::Meta implementation via a \u001b[39m\u001b[33m\"\u001b[39m\n\u001b[32m (...)\u001b[39m\u001b[32m 56\u001b[39m \u001b[33mf\u001b[39m\u001b[33m\"\u001b[39m\u001b[33mregister_fake.\u001b[39m\u001b[33m\"\u001b[39m\n\u001b[32m 57\u001b[39m )\n\u001b[32m 59\u001b[39m \u001b[38;5;28;01mif\u001b[39;00m torch._C._dispatch_has_kernel_for_dispatch_key(\n\u001b[32m 60\u001b[39m \u001b[38;5;28mself\u001b[39m.qualname, \u001b[33m\"\u001b[39m\u001b[33mCompositeImplicitAutograd\u001b[39m\u001b[33m\"\u001b[39m\n\u001b[32m 61\u001b[39m ):\n",
|
|
"\u001b[31mRuntimeError\u001b[39m: operator torchvision::nms does not exist"
|
|
]
|
|
}
|
|
],
|
|
"source": [
|
|
"# no torch no example womp womp\n",
|
|
"import torch\n",
|
|
"import torch.nn as nn\n",
|
|
"import torch.nn.functional as F\n",
|
|
"from torch.utils.data import DataLoader\n",
|
|
"from torchvision import datasets, transforms\n",
|
|
"import numpy as np\n",
|
|
"import matplotlib.pyplot as plt\n",
|
|
"import matplotlib.patches as patches\n",
|
|
"from pathlib import Path\n",
|
|
"import warnings\n",
|
|
"warnings.filterwarnings('ignore')\n",
|
|
"\n",
|
|
"# Import utilities from HTB Evasion Library\n",
|
|
"from htb_ai_library.utils import (\n",
|
|
" set_reproducibility,\n",
|
|
" save_model,\n",
|
|
" load_model,\n",
|
|
" HTB_GREEN,\n",
|
|
" NODE_BLACK,\n",
|
|
" HACKER_GREY,\n",
|
|
" WHITE,\n",
|
|
" AZURE,\n",
|
|
" NUGGET_YELLOW,\n",
|
|
" MALWARE_RED,\n",
|
|
" VIVID_PURPLE,\n",
|
|
" AQUAMARINE,\n",
|
|
")\n",
|
|
"from htb_ai_library.data import get_mnist_loaders\n",
|
|
"from htb_ai_library.models import MNISTClassifierWithDropout\n",
|
|
"from htb_ai_library.training import train_model, evaluate_accuracy\n",
|
|
"from htb_ai_library.visualization import use_htb_style\n",
|
|
"\n",
|
|
"# Apply HTB theme globally to all plots\n",
|
|
"use_htb_style()\n",
|
|
"\n",
|
|
"# Set reproducibility\n",
|
|
"set_reproducibility(1337)\n",
|
|
"\n",
|
|
"# Configure device\n",
|
|
"device = torch.device(\"cuda\" if torch.cuda.is_available() else \"cpu\")\n",
|
|
"print(f\"Using device: {device}\")\n",
|
|
"if device.type == \"cuda\":\n",
|
|
" print(f\"GPU: {torch.cuda.get_device_name(0)}\")"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 2,
|
|
"id": "dfd799cc-157a-4584-89fd-bb0635d0c721",
|
|
"metadata": {},
|
|
"outputs": [
|
|
{
|
|
"ename": "NameError",
|
|
"evalue": "name 'get_mnist_loaders' is not defined",
|
|
"output_type": "error",
|
|
"traceback": [
|
|
"\u001b[31m---------------------------------------------------------------------------\u001b[39m",
|
|
"\u001b[31mNameError\u001b[39m Traceback (most recent call last)",
|
|
"\u001b[36mCell\u001b[39m\u001b[36m \u001b[39m\u001b[32mIn[2]\u001b[39m\u001b[32m, line 2\u001b[39m\n\u001b[32m 1\u001b[39m \u001b[38;5;66;03m# Get data loaders using library function\u001b[39;00m\n\u001b[32m----> \u001b[39m\u001b[32m2\u001b[39m train_loader, test_loader = \u001b[43mget_mnist_loaders\u001b[49m(batch_size=\u001b[32m128\u001b[39m)\n\u001b[32m 3\u001b[39m \u001b[38;5;28mprint\u001b[39m(\u001b[33mf\u001b[39m\u001b[33m\"\u001b[39m\u001b[33mTraining samples: \u001b[39m\u001b[38;5;132;01m{\u001b[39;00m\u001b[38;5;28mlen\u001b[39m(train_loader.dataset)\u001b[38;5;132;01m}\u001b[39;00m\u001b[33m\"\u001b[39m)\n\u001b[32m 4\u001b[39m \u001b[38;5;28mprint\u001b[39m(\u001b[33mf\u001b[39m\u001b[33m\"\u001b[39m\u001b[33mTest samples: \u001b[39m\u001b[38;5;132;01m{\u001b[39;00m\u001b[38;5;28mlen\u001b[39m(test_loader.dataset)\u001b[38;5;132;01m}\u001b[39;00m\u001b[33m\"\u001b[39m)\n",
|
|
"\u001b[31mNameError\u001b[39m: name 'get_mnist_loaders' is not defined"
|
|
]
|
|
}
|
|
],
|
|
"source": [
|
|
"# Get data loaders using library function\n",
|
|
"train_loader, test_loader = get_mnist_loaders(batch_size=128)\n",
|
|
"print(f\"Training samples: {len(train_loader.dataset)}\")\n",
|
|
"print(f\"Test samples: {len(test_loader.dataset)}\")\n",
|
|
"\n",
|
|
"# Create output directory for saving models and results\n",
|
|
"output_dir = Path(\"output\")\n",
|
|
"output_dir.mkdir(exist_ok=True)\n",
|
|
"\n",
|
|
"# Define model checkpoint path in output directory\n",
|
|
"model_path = output_dir / \"mnist_target.pth\"\n",
|
|
"\n",
|
|
"# Initialize model using MNISTClassifierWithDropout from library\n",
|
|
"model = MNISTClassifierWithDropout(num_classes=10).to(device)\n",
|
|
"\n",
|
|
"# Check if trained model exists, otherwise train from scratch\n",
|
|
"if model_path.exists():\n",
|
|
" print(f\"\\nLoading existing model from {model_path}\")\n",
|
|
" model = load_model(model, model_path, device)\n",
|
|
"else:\n",
|
|
" print(f\"\\nNo existing model found. Training new model...\")\n",
|
|
" model = train_model(model, train_loader, test_loader, epochs=5, device=device)\n",
|
|
" print(f\"Saving trained model to {model_path}\")\n",
|
|
" save_model(model, model_path)\n",
|
|
"\n",
|
|
"# Evaluate the trained model\n",
|
|
"accuracy = evaluate_accuracy(model, test_loader, device)\n",
|
|
"print(f\"\\nTest accuracy: {accuracy:.2f}%\")\n"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"id": "ba8f35e6-554b-407e-9e42-07845dc7fd89",
|
|
"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": "3.11.6"
|
|
}
|
|
},
|
|
"nbformat": 4,
|
|
"nbformat_minor": 5
|
|
}
|