{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "c0cffe24-dbf9-4350-92a3-28eb98bcbf4a",
   "metadata": {
    "scrolled": true
   },
   "source": [
    "# Effect of input representations on learning Jupyter Notebook\n",
    "### ICML 2025 Workshop on Methods and Opportunities at Small Scale\n",
    "### May, 2025"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "14ff46aa-29e1-424e-b166-fb16a7f48ec3",
   "metadata": {},
   "source": [
    "This is the accompanying tutorial notebook for the paper *Permutations as a testbed for studying the effect of input representations on learning*. In this notebook, we explain how to generate different representations of permutations, as well as how to compute various statistics on this permutation data for use in downstream classification tasks. We provide the code to generate two example permutation datasets and show how to train a simple MLP classification model on the data. The MLP model performance between the two generated datasets are very different, supporting our interest in the impact that data representation choice can have on learning\n",
    "outcomes.\n",
    "\n",
    "In explaining the specific representations and statistics used to generate results in the paper, we aim to make permutation datasets more readily suited for continuing research in this direction. "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2973f84b-ff0c-431d-bfb7-a4e810b2c247",
   "metadata": {},
   "source": [
    "## Installation:\n",
    "\n",
    "Since SageMath is not pip installable, we need to use conda. The cells below assume you are running this in google colab. \n",
    "\n",
    "Alternatively, to run this notebook locally using a conda environment:\n",
    "```\n",
    "conda create -n permutation sage jupyter ipykernel python=3.11\n",
    "conda activate permutation\n",
    "python -m pip install torch tqdm\n",
    "python -m ipykernel install --user --name permutation --display-name permutation\n",
    "```"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "079d6c42-c0bf-4263-8788-01e0672010b0",
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "#This cell took around 12 minutes to execute\n",
    "!wget -q https://repo.continuum.io/miniconda/Miniconda3-latest-Linux-x86_64.sh\n",
    "!bash Miniconda3-latest-Linux-x86_64.sh -b -p /content/miniconda\n",
    "\n",
    "import os\n",
    "os.environ['PATH'] = \"/content/miniconda/bin:\" + os.environ['PATH']\n",
    "\n",
    "!conda config --add channels conda-forge --quiet\n",
    "!conda install sage python=3.11 -y --quiet"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "68b097b9-d3d1-4f74-9491-272649d907fe",
   "metadata": {},
   "outputs": [],
   "source": [
    "import sys\n",
    "sage_site_packages = '/content/miniconda/lib/python3.11/site-packages'\n",
    "if sage_site_packages not in sys.path:\n",
    "    sys.path.insert(0, sage_site_packages)\n",
    "\n",
    "from sage.all import *"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "id": "00e3810f-8a08-4aa0-bf9a-692b04457afd",
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "# Import relevant packages\n",
    "import numpy as np\n",
    "import math\n",
    "import os\n",
    "import pickle\n",
    "from tqdm import tqdm\n",
    "\n",
    "import torch\n",
    "import torch.nn as nn\n",
    "from torch.utils.data import DataLoader, TensorDataset\n",
    "\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e5b66895-2d10-470a-b5d1-ade3abcb042a",
   "metadata": {},
   "source": [
    "# How to generate different representations of permutations"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ccacca48-fa13-4acd-bf0d-db7ca9d51ca5",
   "metadata": {},
   "source": [
    "We begin with an example permutation in one-line notation."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "id": "6a8bb95b-9344-433a-b4d0-d26f37d5bdcd",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Original permutation (one-line notation): [2, 5, 3, 4, 1]\n"
     ]
    }
   ],
   "source": [
    "x = Permutation([2,5,3,4,1]) \n",
    "print(f'Original permutation (one-line notation): {x}')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e766a933-d163-46b5-a2f7-eb53c28653eb",
   "metadata": {},
   "source": [
    "### Permutation matrix\n",
    "The permutation matrix representation of a permutation is the one-hot encoding of the one line notation, so the $(i, j)$th entry of the permutation matrix is 1 precisely when $\\sigma(i) = j$, and 0 otherwise. The permutation matrix representation of $2 4 3 1$ is:\n",
    "$$\n",
    "\\begin{pmatrix}\n",
    "0 & 1 & 0 & 0 \\\\\n",
    "0 & 0 & 0 & 1 \\\\\n",
    "0 & 0 & 1 & 0 \\\\\n",
    "1 & 0 & 0 & 0\n",
    "\\end{pmatrix}\n",
    "$$"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "040d8002-da1e-4026-a393-d749f38a0ccf",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0 0 0 0 1]\n",
      "[1 0 0 0 0]\n",
      "[0 0 1 0 0]\n",
      "[0 0 0 1 0]\n",
      "[0 1 0 0 0]\n"
     ]
    }
   ],
   "source": [
    "permutation_matrix = x.to_matrix()\n",
    "print(permutation_matrix)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3cac2165-9afa-47bc-aae9-1533d827c895",
   "metadata": {},
   "source": [
    "### Inversion vector\n",
    "The inversion vector of a permutation is the vector where the $i$th entry gives the number of $j > i$ that appear before $i$ in the permutation. The inversion vector of the permutation 2 5 3 4 1 is $(4, 0, 1, 1, 0)$."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "3d9fc41b-e465-4e67-ba21-8f6fa7bbf262",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[4, 0, 1, 1, 0]\n"
     ]
    }
   ],
   "source": [
    "inversion_vector = x.to_inversion_vector()\n",
    "print(inversion_vector)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "29ddb0b9-75f0-476d-a78f-4c64f7b4ba81",
   "metadata": {},
   "source": [
    "### Lehmer cocode\n",
    "The Lehmer cocode is also based on inversions, but the $i$th entry is the number of $j < i$ with $\\sigma(j) > \\sigma(i)$. The Lehmer cocode of the permutation is 2 5 3 4 1 is $(0, 0, 1, 1, 4)$. Note that applying the permutation $\\sigma$ to the Lehmer cocode gives the inversion vector. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "7a35b0fb-462a-41db-a778-0e9bd3820304",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0, 0, 1, 1, 4]\n"
     ]
    }
   ],
   "source": [
    "lehmer_cocode = x.to_lehmer_cocode()\n",
    "print(lehmer_cocode)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6bbd57c3-f6d1-4534-b6a4-0c960c5a0e55",
   "metadata": {},
   "source": [
    "### Major code \n",
    "The *major code* is the vector $(m_1-m_2, m_2 - m_3, \\ldots, m_n)$, where $m_i$ is the sum of all of the descents of the permutation (also called the major index) obtained by erasing letters smaller than $i$ from the permutation. For example, for the permutation $2 5 3 4 1$, $m_1 = 6$, $m_2 =2$, $m_3 = 2$ and $m_4 = m_5 = 0$, so the major code is $(4, 1, 0, 1, 0)$."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "ef07fbb0-6b96-46a3-9ec6-bd0517bbf10a",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[4, 1, 0, 1, 0]\n"
     ]
    }
   ],
   "source": [
    "major_code = x.to_major_code()\n",
    "print(major_code)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b8a33bab-e037-4cde-a613-708aa911389c",
   "metadata": {},
   "source": [
    "### Lexicographically minimal reduced word\n",
    "Finally, we also represent permutations using the reduced word for the permutation that is smallest in lexicographic order. We pad with zeros so that the reduced words are all the same length. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 96,
   "id": "2df44a8e-7b15-4d75-b5e1-9ddd034abf59",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[1 2 3 4 3 2 0 0 0 0]\n"
     ]
    }
   ],
   "source": [
    "def compute_reduced_word_lexmin(x):\n",
    "    n = max(x)\n",
    "    L = Permutation(x).reduced_word_lexmin()\n",
    "    if len(L) < math.comb(n, 2):\n",
    "        newL = L + [0]*(math.comb(n, 2) - len(L))\n",
    "        return np.array(newL)\n",
    "    else:\n",
    "        return np.array(L)\n",
    "\n",
    "reduced_word_lexmin = compute_reduced_word_lexmin(x)\n",
    "print(reduced_word_lexmin)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f2ac7bb0-14ac-4aab-9a26-d6807144a4a1",
   "metadata": {},
   "source": [
    "# How to compute statistics on permutation data"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a1b05041-166e-4bfb-9808-79d89849431b",
   "metadata": {},
   "source": [
    "The permutations module from SAGE math is very useful to quickly compute statistics of permutations. Start with an example permutation, x.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "id": "ee0650a6-2e8c-4915-93f2-abd07d57b5a9",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Original permutation (one-line notation): [2, 5, 3, 4, 1]\n"
     ]
    }
   ],
   "source": [
    "print(f'Original permutation (one-line notation): {x}')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ef4b2dde-df87-49fb-8167-5c2c9321d0a3",
   "metadata": {},
   "source": [
    "### Length\n",
    "The length (or Coxeter length) of a permutation is the number of inversions of the permutation, or, equivalently, the length of a reduced word for the permutation. \n",
    "Note that the length of the permutation is simply the sum of the inversion vector or the Lehmer cocode."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "id": "580acc56-a84a-4082-a73d-9600443c8a66",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "6\n"
     ]
    }
   ],
   "source": [
    "length = x.length()\n",
    "print(length)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0becd76c-36aa-4265-87ea-fbe360fe7ee4",
   "metadata": {},
   "source": [
    "### Permutation parity\n",
    "The parity of a permutation is the even/oddness of its length. A permutation is even if and only if its length is an even number. \n",
    "Note that the parity of a permutation can be computed from the inversion vector or the Lehmer cocode by computing the parity of its sum. \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "id": "b52571b4-fc3c-4d85-abf8-adf205492142",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "True\n"
     ]
    }
   ],
   "source": [
    "parity = x.is_even()\n",
    "print(parity)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4bf4fe21-31e9-48b8-9634-7e96c62b80da",
   "metadata": {},
   "source": [
    "### Major index\n",
    "The major index of a permutation is the sum of all of the descents. \n",
    "Note that the major index is the sum of the major code. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "id": "ac7b3a48-a6ab-47d6-9d6a-8fcf97fc81e7",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "6\n"
     ]
    }
   ],
   "source": [
    "major_index = x.major_index()\n",
    "print(major_index)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d1494f0b-8375-4dda-9a26-c85358df03cb",
   "metadata": {},
   "source": [
    "### Number of descents"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "id": "171db60c-d0c9-4351-9744-5f06b3af6f1f",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2\n"
     ]
    }
   ],
   "source": [
    "num_descents = x.number_of_descents()\n",
    "print(num_descents)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "37672d9e-a271-4ce3-a128-ecd53a457e44",
   "metadata": {},
   "source": [
    "### Number of fixed points\n",
    "The number of fixed points in a permutation are the number of values $i$ such that $\\sigma(i) = i$.\n",
    "Note that the number of fixed points in the permutation is the trace of the permutation matrix. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "id": "d485477f-678c-4e15-9cd9-6f3e9f7295f9",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2\n"
     ]
    }
   ],
   "source": [
    "num_fixed_points = len(x.fixed_points())\n",
    "print(num_fixed_points)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f58918e1-488c-48d7-9ad7-2631af8a369e",
   "metadata": {},
   "source": [
    "### Number of peaks\n",
    "A peak of a permutation is an integer $i$ such that $\\sigma(i-1) < \\sigma(i)$ and $\\sigma(i) > \\sigma(i+1)$.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "id": "3a120dca-20cf-4c62-8995-4b76b6d1c46f",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2\n"
     ]
    }
   ],
   "source": [
    "num_peaks = x.number_of_peaks()\n",
    "print(num_peaks)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b348e6da-cb4d-4b6b-8f00-33a65411b691",
   "metadata": {},
   "source": [
    "### Longest increasing subsequence length\n",
    "The longest increasing subsequence of $\\sigma$ is the longest sequence $i_{1} < i_{2} < \\cdots < i_{n}$ with $\\sigma_{i_{1}} < \\sigma_{i_{2}} < \\cdots < \\sigma_{i_{n}}$. \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "id": "586e9ede-7f6a-41a5-983e-7dd35e1e7106",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "3\n"
     ]
    }
   ],
   "source": [
    "longest_inc_subsequence_len = x.longest_increasing_subsequence_length()\n",
    "print(longest_inc_subsequence_len)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1ce3a835-6e4f-4780-9ffb-9e9deb5d02d2",
   "metadata": {},
   "source": [
    "### Number of nestings\n",
    "A nesting of a permutation is a pair $(i, j)$ such that $j < i < \\sigma(i) < \\sigma(j)$ or $\\sigma(j) < \\sigma(i) \\leq i < j$. This definition comes from the permutation diagram.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "id": "217a2287-0cf0-4e01-b453-cd6ad7a0093e",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2\n"
     ]
    }
   ],
   "source": [
    "def count_nestings(pi):\n",
    "    '''\n",
    "    Code from https://www.findstat.org/StatisticsDatabase/St000223/\n",
    "    '''\n",
    "    ne = 0\n",
    "    n = len(pi)\n",
    "    for i in range(n):\n",
    "        for j in range(n):\n",
    "            if (j < i < pi[i] < pi[j] or pi[j] < pi[i] <= i < j):\n",
    "                ne = ne+1\n",
    "    return ne\n",
    "\n",
    "number_nestings = count_nestings(x)\n",
    "print(number_nestings)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f2f03805-1516-4d68-96f8-cfbcf841e0f5",
   "metadata": {},
   "source": [
    "### Order\n",
    "The order of a permutation is the smallest positive integer $m$ so that applying the permutation $m$ times is equivalent to the identity permutation. The order of a permutation is the least common multiple of the lengths of its cycles. \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "id": "4aabfd18-189e-43c9-a95f-8f2a7943f13e",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "3\n"
     ]
    }
   ],
   "source": [
    "def compute_order(x):\n",
    "    return lcm(x.cycle_type())\n",
    "\n",
    "order = compute_order(x)\n",
    "print(order)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "af3620fc-76ce-4477-8f29-2536d7b0a161",
   "metadata": {},
   "source": [
    "### Number of cycles\n",
    "The number of cycles is the number of cycles in the representation of the permutation as a product of disjoint cycles. (Including cycles of length 1). \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "id": "43fbdcd8-af3e-4058-903b-590dfefe9f10",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "3\n"
     ]
    }
   ],
   "source": [
    "def count_cycles(x):\n",
    "    return len(x.cycle_type())\n",
    "\n",
    "num_cycles = count_cycles(x)\n",
    "print(num_cycles)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9b7004b7-886a-4831-a97c-acc6909584f8",
   "metadata": {},
   "source": [
    "### Cardinality of the support\n",
    "The support of a permutation are the indices that appear in a reduced word of that permutation (which are independent of the reduced word chosen). The cardinality of the support is the number of distinct indices in the support.  \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "id": "eec1ed58-3f34-4f3c-a904-07b17e02e4b3",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "4\n"
     ]
    }
   ],
   "source": [
    "def compute_support_cardinality(x):\n",
    "    '''\n",
    "    Code from https://www.findstat.org/StatisticsDatabase/St000019/\n",
    "    '''\n",
    "    return len(set(x.reduced_word()))\n",
    "\n",
    "support_cardinality = compute_support_cardinality(x)\n",
    "print(support_cardinality)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f0505a3a-56b4-463d-8f4c-cb3667e46d76",
   "metadata": {},
   "source": [
    "### Number of stack sorts\n",
    "The number of stack sorts needed to sort a permutation. A permutation is *stack sortable* if it can be sorted by the following algorithm: \n",
    "\n",
    "- Initialize an empty stack and empty final output\n",
    "- For each element $x$ in the input permutation: (a) while the stack is nonempty and $x$ is larger than the last item added to the stack, pop this item from the stack to the output (b) put $x$ onto the stack\n",
    "- If the stack is nonempty, add remaining stack to the output. \n",
    "\n",
    "When we apply this algorithm to the permutation $2 5 3 4 1$, we get the permutation $2 3 1 4 5$. After two more stack sorts we get the identity permutation, so the number of stack sorts needed to sort the permutation is 3. \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "id": "f5581a23-7a70-43c4-992c-e778f0025f45",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "3\n"
     ]
    }
   ],
   "source": [
    "def stack_sort(x):\n",
    "    '''\n",
    "    Code from https://www.findstat.org/StatisticsDatabase/St000028/\n",
    "    '''\n",
    "    stack = []\n",
    "    result = []\n",
    "    for e in x:\n",
    "        while stack and e > stack[-1]:\n",
    "            result += [stack.pop()]\n",
    "        stack.append(e)\n",
    "    while stack:\n",
    "        result += [stack.pop()]\n",
    "    return Permutation(result)\n",
    "\n",
    "def count_stack_sorts(x):\n",
    "    '''\n",
    "    Code from https://www.findstat.org/StatisticsDatabase/St000028/\n",
    "    '''\n",
    "    i = 0\n",
    "    id = Permutations(x.size()).identity()\n",
    "    while x != id:\n",
    "        x = stack_sort(x)\n",
    "        i += 1\n",
    "    return i\n",
    "\n",
    "num_stack_sorts = count_stack_sorts(x)\n",
    "print(num_stack_sorts)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "611c4e54-4ac7-4d1d-bd6e-d22ddef0aef0",
   "metadata": {},
   "source": [
    "### Further reading"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b407eea4-9a5c-442c-b1ee-ff8636e7dabd",
   "metadata": {},
   "source": [
    "For those interested in exploring more permutation statistics and representations, please see:\n",
    "- [SageMath Permutation Documentation](https://doc.sagemath.org/html/en/reference/combinat/sage/combinat/permutation.html): Comprehensive documentation of permtuation methods in SageMath \n",
    "- [FindStat combinatorial database](https://www.findstat.org/StatisticsDatabase/): A database with over 400 permutation statistics with code implementatison and references: "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c488049b-4aca-4501-9568-e83beabae941",
   "metadata": {},
   "source": [
    "# How to train a MLP to classify permutations by a specified statistic"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f0b41079-480b-48be-b45f-f7dbf4785d57",
   "metadata": {},
   "source": [
    "First, we select a permutation representation, a statistic to classify by, and a value of $N$."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 99,
   "id": "55e7f3db-30fc-485e-9cd6-4ebd78b6c3b2",
   "metadata": {},
   "outputs": [],
   "source": [
    "representation = 'inversion_vector'\n",
    "statistic = 'major_index'\n",
    "N = 7"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d9e70b1c-e4d7-48ff-966e-ef8ad5e782f1",
   "metadata": {},
   "source": [
    "Now, we generate a dataset of permutation samples from $S_{N}$."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a1040eb8-6d0a-4c48-8b8e-2ba5e5ece8f3",
   "metadata": {},
   "source": [
    "Define a function to convert the representations into one-hot-encoded and then flattened vectors."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "id": "4ca51b7b-e79a-4d1a-b837-94c1c1c6e98e",
   "metadata": {},
   "outputs": [],
   "source": [
    "def one_hot_encode(sequences_torch):\n",
    "    \"\"\"\n",
    "    One-hot encodes the elements for a list of permutations.\n",
    "\n",
    "    Args:\n",
    "        sequences_torch (list): A list of torch vectors of permutations to encode.\n",
    "        \n",
    "    Returns:\n",
    "        list: a list of one-hot encoded representations of the permutations flattened in the input vector.\n",
    "    \"\"\"\n",
    "\n",
    "    # Get the maximum value of all of the entries\n",
    "    if sequences_torch[0].ndim == 1:\n",
    "        max_val = float('-inf')  # Initialize with negative infinity\n",
    "        for seq in sequences_torch:\n",
    "            for entry in seq:\n",
    "                if entry > max_val:\n",
    "                    max_val = entry\n",
    "\n",
    "    # Convert tensor to numpy for processing if needed\n",
    "    if not isinstance(sequences_torch, np.ndarray):\n",
    "        sequences_torch = sequences_torch.numpy()\n",
    "    else:\n",
    "        sequences_torch = sequences_torch\n",
    "        \n",
    "    permutations_one_hot_encoded = []\n",
    "    for permutation in sequences_torch:\n",
    "        if np.array(permutation).ndim == 1:\n",
    "            \n",
    "            unique_elements = np.unique(permutation)\n",
    "            encoding_matrix = np.eye(max_val+1)\n",
    "            element_to_index = {element: index for index, element in enumerate(np.array(range(max_val+1)))}\n",
    "            encoded_vector = np.array([encoding_matrix[element_to_index[element]] for element in permutation]).flatten()\n",
    "            \n",
    "        elif permutation.ndim == 2:\n",
    "            encoded_vector = permutation.flatten()\n",
    "            \n",
    "        permutations_one_hot_encoded.append(encoded_vector)\n",
    "        \n",
    "    return torch.tensor(np.array(permutations_one_hot_encoded))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "id": "2d6a3892-709e-43f0-9e4b-c1d22c9788d4",
   "metadata": {},
   "outputs": [],
   "source": [
    "S = Permutations(N)\n",
    "data_lst = [np.array(x.to_inversion_vector()) for x in S]\n",
    "labels_lst = [np.array(x.major_index()) for x in S]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "id": "b5acd289-09f3-482f-9790-bbb4b13b9358",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "DATASET # of classes is: 22\n"
     ]
    }
   ],
   "source": [
    "sequences_torch = torch.from_numpy(np.array(data_lst))\n",
    "labels_torch = torch.from_numpy(np.array(labels_lst)).type(torch.LongTensor)\n",
    "\n",
    "output_size = max(labels_torch)+1\n",
    "print(f\"DATASET # of classes is: {output_size}\")\n",
    "\n",
    "# Use one hot encoding for all permutation inputs\n",
    "sequences_torch = one_hot_encode(sequences_torch)\n",
    "input_size = sequences_torch[0].shape[0]"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4b68c96d-71c1-47e0-833c-e941c5e12761",
   "metadata": {},
   "source": [
    "Do a train test split and define tensor datasets using PyTorch."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "id": "6f08b725-43d8-4acf-bfc9-4b782f26420c",
   "metadata": {},
   "outputs": [],
   "source": [
    "np.random.seed(42)\n",
    "torch.manual_seed(42)\n",
    "\n",
    "split = 0.7\n",
    "ds_size = int(len(sequences_torch))\n",
    "random_idx = list(range(len(sequences_torch)))\n",
    "np.random.shuffle(random_idx)\n",
    "\n",
    "sequences_torch_train = sequences_torch[random_idx][:math.ceil(ds_size*split)].float()\n",
    "labels_torch_train = labels_torch[random_idx][:math.ceil(ds_size*split)]\n",
    "\n",
    "sequences_torch_test = sequences_torch[random_idx][math.ceil(ds_size*split):].float()\n",
    "labels_torch_test = labels_torch[random_idx][math.ceil(ds_size*split):]\n",
    "\n",
    "trainset = torch.utils.data.TensorDataset(sequences_torch_train, labels_torch_train)\n",
    "testset = torch.utils.data.TensorDataset(sequences_torch_test, labels_torch_test)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "817d68b0-cd30-4b6b-bd0e-3d5ef70f2871",
   "metadata": {},
   "source": [
    "Specify model parameters:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "id": "44e81c1e-30ac-47bb-8c8d-455bcf18df3d",
   "metadata": {},
   "outputs": [],
   "source": [
    "device = 'cuda'\n",
    "BATCH_SIZE = 32\n",
    "LR = 0.001\n",
    "DROPOUT = 0.1\n",
    "WEIGHT_DECAY = 0.0001\n",
    "EPOCHS = 200"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1fed88c4-3e67-4c5d-909e-320675a645f7",
   "metadata": {},
   "source": [
    "Now we define the MLP model architecture."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "id": "e1846e03-d786-4095-a5ba-c55ff57c53bd",
   "metadata": {},
   "outputs": [],
   "source": [
    "class FeedForward(nn.Module):\n",
    "    def __init__(self, \n",
    "                 dimensions,\n",
    "                 num_classes,\n",
    "                 dropout_rate = 0.0):\n",
    "        super().__init__()\n",
    "        self.linear_layers = nn.ModuleList([ nn.Linear(dimensions[i], dimensions[i+1]) for i in range(len(dimensions)-1) ]\n",
    "                                            )\n",
    "        self.classification = nn.Linear(dimensions[-1], num_classes)\n",
    "        self.relu = nn.ReLU()\n",
    "\n",
    "        self.drop1 = nn.Dropout(dropout_rate)\n",
    "        self.dropout_rate = dropout_rate\n",
    "    def forward(self, x):\n",
    "        for layer in self.linear_layers:\n",
    "            x = self.relu(layer(x) )\n",
    "            x = self.drop1(x)\n",
    "\n",
    "        x = self.classification(x)\n",
    "\n",
    "        return x"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8a1c5c96-7352-43cc-8160-f447a47d82dd",
   "metadata": {},
   "source": [
    "Define functions to train the MLP."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 88,
   "id": "bfb6898d-2be0-481c-842d-15b9e28d228f",
   "metadata": {},
   "outputs": [],
   "source": [
    "def evaluate_one_epoch(model, testloader, criterion, device):\n",
    "    total_loss = 0\n",
    "    total_correct = 0\n",
    "    total_samples = 0\n",
    "    all_preds = []\n",
    "    all_labels = []\n",
    "    \n",
    "    with torch.no_grad():\n",
    "        for sequences, labels in testloader:\n",
    "            sequences, labels = sequences.float().to(device), labels.to(device)\n",
    "            outputs = model(sequences)\n",
    "            _, predicted = outputs.max(1)\n",
    "            \n",
    "            loss = criterion(outputs, labels)\n",
    "           \n",
    "            total_correct += predicted.eq(labels).sum().item()\n",
    "            total_samples += labels.size(0)\n",
    "            total_loss += loss.item() * labels.size(0)\n",
    "\n",
    "            # Collecting predictions and labels\n",
    "            all_preds.extend(predicted.cpu().numpy())\n",
    "            all_labels.extend(labels.cpu().numpy())\n",
    "\n",
    "    accuracy = total_correct / total_samples\n",
    "    avg_loss = total_loss / total_samples\n",
    "\n",
    "    return avg_loss, accuracy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 89,
   "id": "1ef2e3ce-04c8-449b-8ac7-bf791f5d1bf1",
   "metadata": {},
   "outputs": [],
   "source": [
    "def train(model, trainloader, testloader, criterion, optimizer, epochs=100, bs=32, when_to_print=10):\n",
    "    \n",
    "    all_train_loss, all_train_acc, all_test_loss, all_test_acc, all_test_f1 = [],[], [], [], []\n",
    "    best_test_acc = 0\n",
    "    best_epoch = 0\n",
    "    best_weights = None\n",
    "    \n",
    "    for ep in range(epochs):\n",
    "        model.train()\n",
    "        train_loss, train_correct, train_total = 0, 0, 0\n",
    "\n",
    "        for sequences, labels in tqdm(trainloader, desc=f\"Epoch {ep+1}/{epochs}\"):\n",
    "            sequences, labels = sequences.float().to(device), labels.to(device)\n",
    "            optimizer.zero_grad()\n",
    "            output = model(sequences)\n",
    "            loss = criterion(output, labels)\n",
    "            \n",
    "            # backprop and optimize\n",
    "            loss.backward()\n",
    "            optimizer.step()\n",
    "\n",
    "            _, predicted = output.max(1)\n",
    "            train_correct += predicted.eq(labels).sum().item()\n",
    "            train_total += labels.size(0)\n",
    "            train_loss += loss.item() * labels.size(0)\n",
    "\n",
    "        train_acc = train_correct / train_total\n",
    "        avg_train_loss = train_loss / train_total\n",
    "        all_train_acc.append(train_acc)\n",
    "        all_train_loss.append(avg_train_loss)\n",
    "\n",
    "        if ep % when_to_print == 0:\n",
    "            print(f\"Epoch {ep}: Train Acc: {train_acc:.2%}, Train Loss: {avg_train_loss:.4f}\")\n",
    "\n",
    "        model.eval()\n",
    "        avg_test_loss, test_acc = evaluate_one_epoch(model, testloader, criterion, device)\n",
    "\n",
    "        all_test_acc.append(test_acc)\n",
    "        all_test_loss.append(avg_test_loss)\n",
    "\n",
    "        if ep % when_to_print == 0:\n",
    "            print(f\"Test Acc: {test_acc:.2%}, Test Loss: {avg_test_loss:.4f}\")\n",
    "                \n",
    "    return all_train_loss, all_train_acc, all_test_loss, all_test_acc"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ca5d114e-80d8-4133-ba36-818a55fc47a2",
   "metadata": {},
   "source": [
    "Now train the model."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 90,
   "id": "bee7fb69-5c14-4307-b23d-848d95b4a015",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training the MLP for inversion_vector, major_index...\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 1/200: 100%|██████████| 111/111 [00:00<00:00, 471.95it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 0: Train Acc: 9.78%, Train Loss: 2.7268\n",
      "Test Acc: 13.43%, Test Loss: 2.5439\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 2/200: 100%|██████████| 111/111 [00:00<00:00, 474.98it/s]\n",
      "Epoch 3/200: 100%|██████████| 111/111 [00:00<00:00, 480.43it/s]\n",
      "Epoch 4/200: 100%|██████████| 111/111 [00:00<00:00, 480.34it/s]\n",
      "Epoch 5/200: 100%|██████████| 111/111 [00:00<00:00, 481.04it/s]\n",
      "Epoch 6/200: 100%|██████████| 111/111 [00:00<00:00, 481.11it/s]\n",
      "Epoch 7/200: 100%|██████████| 111/111 [00:00<00:00, 479.67it/s]\n",
      "Epoch 8/200: 100%|██████████| 111/111 [00:00<00:00, 479.63it/s]\n",
      "Epoch 9/200: 100%|██████████| 111/111 [00:00<00:00, 480.46it/s]\n",
      "Epoch 10/200: 100%|██████████| 111/111 [00:00<00:00, 475.36it/s]\n",
      "Epoch 11/200: 100%|██████████| 111/111 [00:00<00:00, 479.78it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 10: Train Acc: 40.02%, Train Loss: 1.5018\n",
      "Test Acc: 38.76%, Test Loss: 1.4371\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 12/200: 100%|██████████| 111/111 [00:00<00:00, 479.46it/s]\n",
      "Epoch 13/200: 100%|██████████| 111/111 [00:00<00:00, 478.96it/s]\n",
      "Epoch 14/200: 100%|██████████| 111/111 [00:00<00:00, 479.56it/s]\n",
      "Epoch 15/200: 100%|██████████| 111/111 [00:00<00:00, 440.84it/s]\n",
      "Epoch 16/200: 100%|██████████| 111/111 [00:00<00:00, 479.89it/s]\n",
      "Epoch 17/200: 100%|██████████| 111/111 [00:00<00:00, 479.64it/s]\n",
      "Epoch 18/200: 100%|██████████| 111/111 [00:00<00:00, 475.31it/s]\n",
      "Epoch 19/200: 100%|██████████| 111/111 [00:00<00:00, 479.89it/s]\n",
      "Epoch 20/200: 100%|██████████| 111/111 [00:00<00:00, 480.10it/s]\n",
      "Epoch 21/200: 100%|██████████| 111/111 [00:00<00:00, 480.14it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 20: Train Acc: 64.91%, Train Loss: 0.9382\n",
      "Test Acc: 70.70%, Test Loss: 0.8534\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 22/200: 100%|██████████| 111/111 [00:00<00:00, 479.37it/s]\n",
      "Epoch 23/200: 100%|██████████| 111/111 [00:00<00:00, 479.86it/s]\n",
      "Epoch 24/200: 100%|██████████| 111/111 [00:00<00:00, 478.87it/s]\n",
      "Epoch 25/200: 100%|██████████| 111/111 [00:00<00:00, 474.04it/s]\n",
      "Epoch 26/200: 100%|██████████| 111/111 [00:00<00:00, 475.73it/s]\n",
      "Epoch 27/200: 100%|██████████| 111/111 [00:00<00:00, 480.41it/s]\n",
      "Epoch 28/200: 100%|██████████| 111/111 [00:00<00:00, 480.32it/s]\n",
      "Epoch 29/200: 100%|██████████| 111/111 [00:00<00:00, 479.47it/s]\n",
      "Epoch 30/200: 100%|██████████| 111/111 [00:00<00:00, 479.33it/s]\n",
      "Epoch 31/200: 100%|██████████| 111/111 [00:00<00:00, 479.71it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 30: Train Acc: 78.83%, Train Loss: 0.5929\n",
      "Test Acc: 84.99%, Test Loss: 0.4878\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 32/200: 100%|██████████| 111/111 [00:00<00:00, 479.49it/s]\n",
      "Epoch 33/200: 100%|██████████| 111/111 [00:00<00:00, 476.35it/s]\n",
      "Epoch 34/200: 100%|██████████| 111/111 [00:00<00:00, 475.16it/s]\n",
      "Epoch 35/200: 100%|██████████| 111/111 [00:00<00:00, 479.01it/s]\n",
      "Epoch 36/200: 100%|██████████| 111/111 [00:00<00:00, 478.19it/s]\n",
      "Epoch 37/200: 100%|██████████| 111/111 [00:00<00:00, 479.46it/s]\n",
      "Epoch 38/200: 100%|██████████| 111/111 [00:00<00:00, 479.73it/s]\n",
      "Epoch 39/200: 100%|██████████| 111/111 [00:00<00:00, 479.54it/s]\n",
      "Epoch 40/200: 100%|██████████| 111/111 [00:00<00:00, 478.53it/s]\n",
      "Epoch 41/200: 100%|██████████| 111/111 [00:00<00:00, 479.43it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 40: Train Acc: 85.40%, Train Loss: 0.3956\n",
      "Test Acc: 92.13%, Test Loss: 0.2723\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 42/200: 100%|██████████| 111/111 [00:00<00:00, 475.02it/s]\n",
      "Epoch 43/200: 100%|██████████| 111/111 [00:00<00:00, 479.76it/s]\n",
      "Epoch 44/200: 100%|██████████| 111/111 [00:00<00:00, 479.56it/s]\n",
      "Epoch 45/200: 100%|██████████| 111/111 [00:00<00:00, 479.42it/s]\n",
      "Epoch 46/200: 100%|██████████| 111/111 [00:00<00:00, 479.65it/s]\n",
      "Epoch 47/200: 100%|██████████| 111/111 [00:00<00:00, 479.02it/s]\n",
      "Epoch 48/200: 100%|██████████| 111/111 [00:00<00:00, 479.70it/s]\n",
      "Epoch 49/200: 100%|██████████| 111/111 [00:00<00:00, 479.94it/s]\n",
      "Epoch 50/200: 100%|██████████| 111/111 [00:00<00:00, 474.51it/s]\n",
      "Epoch 51/200: 100%|██████████| 111/111 [00:00<00:00, 479.14it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 50: Train Acc: 90.59%, Train Loss: 0.2746\n",
      "Test Acc: 95.90%, Test Loss: 0.1699\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 52/200: 100%|██████████| 111/111 [00:00<00:00, 479.09it/s]\n",
      "Epoch 53/200: 100%|██████████| 111/111 [00:00<00:00, 440.86it/s]\n",
      "Epoch 54/200: 100%|██████████| 111/111 [00:00<00:00, 478.98it/s]\n",
      "Epoch 55/200: 100%|██████████| 111/111 [00:00<00:00, 479.40it/s]\n",
      "Epoch 56/200: 100%|██████████| 111/111 [00:00<00:00, 478.66it/s]\n",
      "Epoch 57/200: 100%|██████████| 111/111 [00:00<00:00, 479.22it/s]\n",
      "Epoch 58/200: 100%|██████████| 111/111 [00:00<00:00, 474.39it/s]\n",
      "Epoch 59/200: 100%|██████████| 111/111 [00:00<00:00, 479.12it/s]\n",
      "Epoch 60/200: 100%|██████████| 111/111 [00:00<00:00, 478.98it/s]\n",
      "Epoch 61/200: 100%|██████████| 111/111 [00:00<00:00, 479.05it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 60: Train Acc: 93.06%, Train Loss: 0.2121\n",
      "Test Acc: 97.09%, Test Loss: 0.1033\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 62/200: 100%|██████████| 111/111 [00:00<00:00, 473.07it/s]\n",
      "Epoch 63/200: 100%|██████████| 111/111 [00:00<00:00, 479.22it/s]\n",
      "Epoch 64/200: 100%|██████████| 111/111 [00:00<00:00, 479.05it/s]\n",
      "Epoch 65/200: 100%|██████████| 111/111 [00:00<00:00, 478.76it/s]\n",
      "Epoch 66/200: 100%|██████████| 111/111 [00:00<00:00, 478.64it/s]\n",
      "Epoch 67/200: 100%|██████████| 111/111 [00:00<00:00, 473.83it/s]\n",
      "Epoch 68/200: 100%|██████████| 111/111 [00:00<00:00, 478.50it/s]\n",
      "Epoch 69/200: 100%|██████████| 111/111 [00:00<00:00, 479.02it/s]\n",
      "Epoch 70/200: 100%|██████████| 111/111 [00:00<00:00, 478.47it/s]\n",
      "Epoch 71/200: 100%|██████████| 111/111 [00:00<00:00, 478.94it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 70: Train Acc: 94.67%, Train Loss: 0.1761\n",
      "Test Acc: 97.62%, Test Loss: 0.0829\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 72/200: 100%|██████████| 111/111 [00:00<00:00, 478.92it/s]\n",
      "Epoch 73/200: 100%|██████████| 111/111 [00:00<00:00, 477.87it/s]\n",
      "Epoch 74/200: 100%|██████████| 111/111 [00:00<00:00, 475.65it/s]\n",
      "Epoch 75/200: 100%|██████████| 111/111 [00:00<00:00, 474.81it/s]\n",
      "Epoch 76/200: 100%|██████████| 111/111 [00:00<00:00, 478.54it/s]\n",
      "Epoch 77/200: 100%|██████████| 111/111 [00:00<00:00, 478.43it/s]\n",
      "Epoch 78/200: 100%|██████████| 111/111 [00:00<00:00, 478.75it/s]\n",
      "Epoch 79/200: 100%|██████████| 111/111 [00:00<00:00, 477.66it/s]\n",
      "Epoch 80/200: 100%|██████████| 111/111 [00:00<00:00, 478.37it/s]\n",
      "Epoch 81/200: 100%|██████████| 111/111 [00:00<00:00, 479.05it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 80: Train Acc: 95.46%, Train Loss: 0.1339\n",
      "Test Acc: 99.07%, Test Loss: 0.0482\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 82/200: 100%|██████████| 111/111 [00:00<00:00, 478.77it/s]\n",
      "Epoch 83/200: 100%|██████████| 111/111 [00:00<00:00, 474.09it/s]\n",
      "Epoch 84/200: 100%|██████████| 111/111 [00:00<00:00, 479.13it/s]\n",
      "Epoch 85/200: 100%|██████████| 111/111 [00:00<00:00, 438.92it/s]\n",
      "Epoch 86/200: 100%|██████████| 111/111 [00:00<00:00, 478.63it/s]\n",
      "Epoch 87/200: 100%|██████████| 111/111 [00:00<00:00, 478.88it/s]\n",
      "Epoch 88/200: 100%|██████████| 111/111 [00:00<00:00, 479.15it/s]\n",
      "Epoch 89/200: 100%|██████████| 111/111 [00:00<00:00, 443.09it/s]\n",
      "Epoch 90/200: 100%|██████████| 111/111 [00:00<00:00, 478.69it/s]\n",
      "Epoch 91/200: 100%|██████████| 111/111 [00:00<00:00, 474.46it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 90: Train Acc: 96.23%, Train Loss: 0.1195\n",
      "Test Acc: 98.94%, Test Loss: 0.0429\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 92/200: 100%|██████████| 111/111 [00:00<00:00, 478.72it/s]\n",
      "Epoch 93/200: 100%|██████████| 111/111 [00:00<00:00, 478.91it/s]\n",
      "Epoch 94/200: 100%|██████████| 111/111 [00:00<00:00, 479.09it/s]\n",
      "Epoch 95/200: 100%|██████████| 111/111 [00:00<00:00, 478.64it/s]\n",
      "Epoch 96/200: 100%|██████████| 111/111 [00:00<00:00, 479.26it/s]\n",
      "Epoch 97/200: 100%|██████████| 111/111 [00:00<00:00, 478.53it/s]\n",
      "Epoch 98/200: 100%|██████████| 111/111 [00:00<00:00, 478.08it/s]\n",
      "Epoch 99/200: 100%|██████████| 111/111 [00:00<00:00, 468.40it/s]\n",
      "Epoch 100/200: 100%|██████████| 111/111 [00:00<00:00, 477.95it/s]\n",
      "Epoch 101/200: 100%|██████████| 111/111 [00:00<00:00, 478.26it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 100: Train Acc: 96.60%, Train Loss: 0.1102\n",
      "Test Acc: 98.94%, Test Loss: 0.0338\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 102/200: 100%|██████████| 111/111 [00:00<00:00, 478.70it/s]\n",
      "Epoch 103/200: 100%|██████████| 111/111 [00:00<00:00, 479.16it/s]\n",
      "Epoch 104/200: 100%|██████████| 111/111 [00:00<00:00, 478.84it/s]\n",
      "Epoch 105/200: 100%|██████████| 111/111 [00:00<00:00, 478.52it/s]\n",
      "Epoch 106/200: 100%|██████████| 111/111 [00:00<00:00, 479.05it/s]\n",
      "Epoch 107/200: 100%|██████████| 111/111 [00:00<00:00, 471.26it/s]\n",
      "Epoch 108/200: 100%|██████████| 111/111 [00:00<00:00, 478.68it/s]\n",
      "Epoch 109/200: 100%|██████████| 111/111 [00:00<00:00, 478.41it/s]\n",
      "Epoch 110/200: 100%|██████████| 111/111 [00:00<00:00, 478.67it/s]\n",
      "Epoch 111/200: 100%|██████████| 111/111 [00:00<00:00, 478.89it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 110: Train Acc: 97.08%, Train Loss: 0.0950\n",
      "Test Acc: 99.47%, Test Loss: 0.0336\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 112/200: 100%|██████████| 111/111 [00:00<00:00, 478.80it/s]\n",
      "Epoch 113/200: 100%|██████████| 111/111 [00:00<00:00, 477.85it/s]\n",
      "Epoch 114/200: 100%|██████████| 111/111 [00:00<00:00, 478.55it/s]\n",
      "Epoch 115/200: 100%|██████████| 111/111 [00:00<00:00, 473.89it/s]\n",
      "Epoch 116/200: 100%|██████████| 111/111 [00:00<00:00, 478.36it/s]\n",
      "Epoch 117/200: 100%|██████████| 111/111 [00:00<00:00, 478.54it/s]\n",
      "Epoch 118/200: 100%|██████████| 111/111 [00:00<00:00, 478.83it/s]\n",
      "Epoch 119/200: 100%|██████████| 111/111 [00:00<00:00, 478.53it/s]\n",
      "Epoch 120/200: 100%|██████████| 111/111 [00:00<00:00, 478.53it/s]\n",
      "Epoch 121/200: 100%|██████████| 111/111 [00:00<00:00, 478.18it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 120: Train Acc: 97.82%, Train Loss: 0.0751\n",
      "Test Acc: 99.47%, Test Loss: 0.0266\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 122/200: 100%|██████████| 111/111 [00:00<00:00, 478.65it/s]\n",
      "Epoch 123/200: 100%|██████████| 111/111 [00:00<00:00, 473.67it/s]\n",
      "Epoch 124/200: 100%|██████████| 111/111 [00:00<00:00, 478.30it/s]\n",
      "Epoch 125/200: 100%|██████████| 111/111 [00:00<00:00, 478.85it/s]\n",
      "Epoch 126/200: 100%|██████████| 111/111 [00:00<00:00, 439.51it/s]\n",
      "Epoch 127/200: 100%|██████████| 111/111 [00:00<00:00, 477.73it/s]\n",
      "Epoch 128/200: 100%|██████████| 111/111 [00:00<00:00, 478.51it/s]\n",
      "Epoch 129/200: 100%|██████████| 111/111 [00:00<00:00, 478.17it/s]\n",
      "Epoch 130/200: 100%|██████████| 111/111 [00:00<00:00, 478.13it/s]\n",
      "Epoch 131/200: 100%|██████████| 111/111 [00:00<00:00, 473.68it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 130: Train Acc: 97.73%, Train Loss: 0.0755\n",
      "Test Acc: 99.21%, Test Loss: 0.0269\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 132/200: 100%|██████████| 111/111 [00:00<00:00, 477.31it/s]\n",
      "Epoch 133/200: 100%|██████████| 111/111 [00:00<00:00, 477.12it/s]\n",
      "Epoch 134/200: 100%|██████████| 111/111 [00:00<00:00, 477.90it/s]\n",
      "Epoch 135/200: 100%|██████████| 111/111 [00:00<00:00, 478.44it/s]\n",
      "Epoch 136/200: 100%|██████████| 111/111 [00:00<00:00, 474.69it/s]\n",
      "Epoch 137/200: 100%|██████████| 111/111 [00:00<00:00, 477.55it/s]\n",
      "Epoch 138/200: 100%|██████████| 111/111 [00:00<00:00, 477.85it/s]\n",
      "Epoch 139/200: 100%|██████████| 111/111 [00:00<00:00, 473.74it/s]\n",
      "Epoch 140/200: 100%|██████████| 111/111 [00:00<00:00, 478.30it/s]\n",
      "Epoch 141/200: 100%|██████████| 111/111 [00:00<00:00, 477.97it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 140: Train Acc: 97.17%, Train Loss: 0.0797\n",
      "Test Acc: 99.47%, Test Loss: 0.0223\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 142/200: 100%|██████████| 111/111 [00:00<00:00, 477.45it/s]\n",
      "Epoch 143/200: 100%|██████████| 111/111 [00:00<00:00, 477.76it/s]\n",
      "Epoch 144/200: 100%|██████████| 111/111 [00:00<00:00, 473.71it/s]\n",
      "Epoch 145/200: 100%|██████████| 111/111 [00:00<00:00, 477.44it/s]\n",
      "Epoch 146/200: 100%|██████████| 111/111 [00:00<00:00, 476.21it/s]\n",
      "Epoch 147/200: 100%|██████████| 111/111 [00:00<00:00, 473.80it/s]\n",
      "Epoch 148/200: 100%|██████████| 111/111 [00:00<00:00, 478.10it/s]\n",
      "Epoch 149/200: 100%|██████████| 111/111 [00:00<00:00, 477.66it/s]\n",
      "Epoch 150/200: 100%|██████████| 111/111 [00:00<00:00, 477.85it/s]\n",
      "Epoch 151/200: 100%|██████████| 111/111 [00:00<00:00, 478.64it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 150: Train Acc: 98.24%, Train Loss: 0.0647\n",
      "Test Acc: 99.54%, Test Loss: 0.0190\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 152/200: 100%|██████████| 111/111 [00:00<00:00, 478.13it/s]\n",
      "Epoch 153/200: 100%|██████████| 111/111 [00:00<00:00, 478.11it/s]\n",
      "Epoch 154/200: 100%|██████████| 111/111 [00:00<00:00, 477.59it/s]\n",
      "Epoch 155/200: 100%|██████████| 111/111 [00:00<00:00, 473.52it/s]\n",
      "Epoch 156/200: 100%|██████████| 111/111 [00:00<00:00, 478.20it/s]\n",
      "Epoch 157/200: 100%|██████████| 111/111 [00:00<00:00, 478.13it/s]\n",
      "Epoch 158/200: 100%|██████████| 111/111 [00:00<00:00, 478.09it/s]\n",
      "Epoch 159/200: 100%|██████████| 111/111 [00:00<00:00, 477.53it/s]\n",
      "Epoch 160/200: 100%|██████████| 111/111 [00:00<00:00, 477.90it/s]\n",
      "Epoch 161/200: 100%|██████████| 111/111 [00:00<00:00, 477.90it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 160: Train Acc: 98.27%, Train Loss: 0.0626\n",
      "Test Acc: 99.80%, Test Loss: 0.0118\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 162/200: 100%|██████████| 111/111 [00:00<00:00, 478.17it/s]\n",
      "Epoch 163/200: 100%|██████████| 111/111 [00:00<00:00, 435.33it/s]\n",
      "Epoch 164/200: 100%|██████████| 111/111 [00:00<00:00, 477.50it/s]\n",
      "Epoch 165/200: 100%|██████████| 111/111 [00:00<00:00, 477.30it/s]\n",
      "Epoch 166/200: 100%|██████████| 111/111 [00:00<00:00, 477.75it/s]\n",
      "Epoch 167/200: 100%|██████████| 111/111 [00:00<00:00, 478.20it/s]\n",
      "Epoch 168/200: 100%|██████████| 111/111 [00:00<00:00, 478.03it/s]\n",
      "Epoch 169/200: 100%|██████████| 111/111 [00:00<00:00, 478.19it/s]\n",
      "Epoch 170/200: 100%|██████████| 111/111 [00:00<00:00, 478.48it/s]\n",
      "Epoch 171/200: 100%|██████████| 111/111 [00:00<00:00, 473.15it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 170: Train Acc: 98.55%, Train Loss: 0.0491\n",
      "Test Acc: 99.87%, Test Loss: 0.0107\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 172/200: 100%|██████████| 111/111 [00:00<00:00, 477.83it/s]\n",
      "Epoch 173/200: 100%|██████████| 111/111 [00:00<00:00, 471.80it/s]\n",
      "Epoch 174/200: 100%|██████████| 111/111 [00:00<00:00, 477.72it/s]\n",
      "Epoch 175/200: 100%|██████████| 111/111 [00:00<00:00, 478.07it/s]\n",
      "Epoch 176/200: 100%|██████████| 111/111 [00:00<00:00, 476.75it/s]\n",
      "Epoch 177/200: 100%|██████████| 111/111 [00:00<00:00, 477.48it/s]\n",
      "Epoch 178/200: 100%|██████████| 111/111 [00:00<00:00, 477.49it/s]\n",
      "Epoch 179/200: 100%|██████████| 111/111 [00:00<00:00, 473.45it/s]\n",
      "Epoch 180/200: 100%|██████████| 111/111 [00:00<00:00, 478.21it/s]\n",
      "Epoch 181/200: 100%|██████████| 111/111 [00:00<00:00, 474.73it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 180: Train Acc: 97.93%, Train Loss: 0.0650\n",
      "Test Acc: 99.93%, Test Loss: 0.0099\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 182/200: 100%|██████████| 111/111 [00:00<00:00, 478.02it/s]\n",
      "Epoch 183/200: 100%|██████████| 111/111 [00:00<00:00, 477.95it/s]\n",
      "Epoch 184/200: 100%|██████████| 111/111 [00:00<00:00, 478.59it/s]\n",
      "Epoch 185/200: 100%|██████████| 111/111 [00:00<00:00, 477.31it/s]\n",
      "Epoch 186/200: 100%|██████████| 111/111 [00:00<00:00, 477.80it/s]\n",
      "Epoch 187/200: 100%|██████████| 111/111 [00:00<00:00, 473.38it/s]\n",
      "Epoch 188/200: 100%|██████████| 111/111 [00:00<00:00, 477.86it/s]\n",
      "Epoch 189/200: 100%|██████████| 111/111 [00:00<00:00, 477.81it/s]\n",
      "Epoch 190/200: 100%|██████████| 111/111 [00:00<00:00, 478.14it/s]\n",
      "Epoch 191/200: 100%|██████████| 111/111 [00:00<00:00, 478.31it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 190: Train Acc: 98.75%, Train Loss: 0.0377\n",
      "Test Acc: 99.80%, Test Loss: 0.0134\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 192/200: 100%|██████████| 111/111 [00:00<00:00, 477.87it/s]\n",
      "Epoch 193/200: 100%|██████████| 111/111 [00:00<00:00, 477.72it/s]\n",
      "Epoch 194/200: 100%|██████████| 111/111 [00:00<00:00, 477.62it/s]\n",
      "Epoch 195/200: 100%|██████████| 111/111 [00:00<00:00, 473.68it/s]\n",
      "Epoch 196/200: 100%|██████████| 111/111 [00:00<00:00, 478.19it/s]\n",
      "Epoch 197/200: 100%|██████████| 111/111 [00:00<00:00, 477.39it/s]\n",
      "Epoch 198/200: 100%|██████████| 111/111 [00:00<00:00, 478.18it/s]\n",
      "Epoch 199/200: 100%|██████████| 111/111 [00:00<00:00, 477.49it/s]\n",
      "Epoch 200/200: 100%|██████████| 111/111 [00:00<00:00, 477.55it/s]\n"
     ]
    }
   ],
   "source": [
    "print(f'Training the MLP for {representation}, {statistic}...')\n",
    "best_accuracy = 0\n",
    "\n",
    "# Create and train the model with the specified hyperparameters\n",
    "model = FeedForward(dimensions=[input_size, 128, 128, 64], num_classes=output_size, dropout_rate=DROPOUT)\n",
    "model.to(device)\n",
    "criterion = nn.CrossEntropyLoss()\n",
    "optimizer = torch.optim.Adam(model.parameters(), lr=LR, weight_decay=WEIGHT_DECAY)\n",
    "\n",
    "trainloader = DataLoader(trainset, batch_size=BATCH_SIZE, shuffle=True)\n",
    "testloader = DataLoader(testset, batch_size=BATCH_SIZE, shuffle=False)\n",
    "\n",
    "# Train the model\n",
    "all_train_loss, all_train_acc, all_test_loss, all_test_acc = train( model, trainloader, testloader, criterion, \n",
    "                                                                   optimizer, epochs=EPOCHS, bs=BATCH_SIZE, when_to_print=10)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "404c49be-dd8d-4418-8a33-357ab5dabe14",
   "metadata": {},
   "source": [
    "Plot the results by loss and accuracy."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 91,
   "id": "b253bd8d-ebf3-4771-8bd7-59c362df5163",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 800x550 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.style.use(\"seaborn-v0_8\")\n",
    "plt.title(f'Loss curve for representation:{representation},\\n statistic:{statistic},\\n N = {N}')\n",
    "plt.xlabel('EPOCH')\n",
    "plt.ylabel('Loss')\n",
    "plt.plot(all_train_loss, label = 'train data')\n",
    "plt.plot(all_test_loss, label= 'test data')\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 92,
   "id": "74b1167b-e111-4bc2-9bd5-d1c86302717f",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 800x550 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.title(f'Model accuracy for representation:{representation},\\n statistic:{statistic},\\n N = {N}')\n",
    "plt.xlabel('EPOCH')\n",
    "plt.ylabel('Accuracy')\n",
    "plt.plot(all_train_acc, label = 'train data')\n",
    "plt.plot(all_test_acc, label= 'test data')\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "393cbd15-1a8a-434b-bb75-9491ac7bb2fc",
   "metadata": {},
   "source": [
    "### Now train a model using a different input representation\n",
    "We use a permutation dataset of the same permutations (from $S_N$) to classify by major code.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 100,
   "id": "2dcb2436-800c-40ee-aff0-82f2788a7a58",
   "metadata": {},
   "outputs": [],
   "source": [
    "representation = 'reduced_word_lexmin'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 101,
   "id": "a2e26279-64b5-42c3-9de6-48d07dbe9427",
   "metadata": {},
   "outputs": [],
   "source": [
    "S = Permutations(N)\n",
    "data_lst = [np.array(compute_reduced_word_lexmin(x)) for x in S]\n",
    "labels_lst = [np.array(x.major_index()) for x in S]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 104,
   "id": "84e8c114-1072-43f2-a63e-3beb17b26799",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "DATASET # of classes is: 22\n"
     ]
    }
   ],
   "source": [
    "sequences_torch = torch.from_numpy(np.array(data_lst))\n",
    "labels_torch = torch.from_numpy(np.array(labels_lst)).type(torch.LongTensor)\n",
    "\n",
    "output_size = max(labels_torch)+1\n",
    "print(f\"DATASET # of classes is: {output_size}\")\n",
    "\n",
    "# Use one hot encoding for all permutation inputs\n",
    "sequences_torch = one_hot_encode(sequences_torch)\n",
    "input_size = sequences_torch[0].shape[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 105,
   "id": "51874dac-64ed-4814-b1fe-a50bbc87c70b",
   "metadata": {},
   "outputs": [],
   "source": [
    "np.random.seed(42)\n",
    "torch.manual_seed(42)\n",
    "\n",
    "split = 0.7\n",
    "ds_size = int(len(sequences_torch))\n",
    "random_idx = list(range(len(sequences_torch)))\n",
    "np.random.shuffle(random_idx)\n",
    "\n",
    "sequences_torch_train = sequences_torch[random_idx][:math.ceil(ds_size*split)].float()\n",
    "labels_torch_train = labels_torch[random_idx][:math.ceil(ds_size*split)]\n",
    "\n",
    "sequences_torch_test = sequences_torch[random_idx][math.ceil(ds_size*split):].float()\n",
    "labels_torch_test = labels_torch[random_idx][math.ceil(ds_size*split):]\n",
    "\n",
    "trainset = torch.utils.data.TensorDataset(sequences_torch_train, labels_torch_train)\n",
    "testset = torch.utils.data.TensorDataset(sequences_torch_test, labels_torch_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 106,
   "id": "0485bcc2-f44f-4aec-a0e6-493fdf52c1a3",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training the MLP for reduced_word_lexmin, major_index...\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 1/200: 100%|██████████| 111/111 [00:00<00:00, 471.18it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 0: Train Acc: 10.74%, Train Loss: 2.7192\n",
      "Test Acc: 12.10%, Test Loss: 2.5495\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 2/200: 100%|██████████| 111/111 [00:00<00:00, 475.44it/s]\n",
      "Epoch 3/200: 100%|██████████| 111/111 [00:00<00:00, 476.91it/s]\n",
      "Epoch 4/200: 100%|██████████| 111/111 [00:00<00:00, 476.52it/s]\n",
      "Epoch 5/200: 100%|██████████| 111/111 [00:00<00:00, 476.63it/s]\n",
      "Epoch 6/200: 100%|██████████| 111/111 [00:00<00:00, 477.16it/s]\n",
      "Epoch 7/200: 100%|██████████| 111/111 [00:00<00:00, 476.71it/s]\n",
      "Epoch 8/200: 100%|██████████| 111/111 [00:00<00:00, 476.03it/s]\n",
      "Epoch 9/200: 100%|██████████| 111/111 [00:00<00:00, 475.82it/s]\n",
      "Epoch 10/200: 100%|██████████| 111/111 [00:00<00:00, 476.30it/s]\n",
      "Epoch 11/200: 100%|██████████| 111/111 [00:00<00:00, 476.43it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 10: Train Acc: 15.90%, Train Loss: 2.1598\n",
      "Test Acc: 17.26%, Test Loss: 2.2456\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 12/200: 100%|██████████| 111/111 [00:00<00:00, 471.43it/s]\n",
      "Epoch 13/200: 100%|██████████| 111/111 [00:00<00:00, 476.59it/s]\n",
      "Epoch 14/200: 100%|██████████| 111/111 [00:00<00:00, 436.82it/s]\n",
      "Epoch 15/200: 100%|██████████| 111/111 [00:00<00:00, 476.48it/s]\n",
      "Epoch 16/200: 100%|██████████| 111/111 [00:00<00:00, 476.13it/s]\n",
      "Epoch 17/200: 100%|██████████| 111/111 [00:00<00:00, 476.00it/s]\n",
      "Epoch 18/200: 100%|██████████| 111/111 [00:00<00:00, 476.35it/s]\n",
      "Epoch 19/200: 100%|██████████| 111/111 [00:00<00:00, 476.37it/s]\n",
      "Epoch 20/200: 100%|██████████| 111/111 [00:00<00:00, 428.65it/s]\n",
      "Epoch 21/200: 100%|██████████| 111/111 [00:00<00:00, 476.41it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 20: Train Acc: 21.63%, Train Loss: 1.9348\n",
      "Test Acc: 17.06%, Test Loss: 2.1860\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 22/200: 100%|██████████| 111/111 [00:00<00:00, 476.39it/s]\n",
      "Epoch 23/200: 100%|██████████| 111/111 [00:00<00:00, 475.79it/s]\n",
      "Epoch 24/200: 100%|██████████| 111/111 [00:00<00:00, 475.03it/s]\n",
      "Epoch 25/200: 100%|██████████| 111/111 [00:00<00:00, 476.46it/s]\n",
      "Epoch 26/200: 100%|██████████| 111/111 [00:00<00:00, 476.31it/s]\n",
      "Epoch 27/200: 100%|██████████| 111/111 [00:00<00:00, 475.67it/s]\n",
      "Epoch 28/200: 100%|██████████| 111/111 [00:00<00:00, 471.54it/s]\n",
      "Epoch 29/200: 100%|██████████| 111/111 [00:00<00:00, 475.37it/s]\n",
      "Epoch 30/200: 100%|██████████| 111/111 [00:00<00:00, 476.25it/s]\n",
      "Epoch 31/200: 100%|██████████| 111/111 [00:00<00:00, 475.96it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 30: Train Acc: 29.25%, Train Loss: 1.7174\n",
      "Test Acc: 18.39%, Test Loss: 2.1291\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 32/200: 100%|██████████| 111/111 [00:00<00:00, 475.95it/s]\n",
      "Epoch 33/200: 100%|██████████| 111/111 [00:00<00:00, 476.56it/s]\n",
      "Epoch 34/200: 100%|██████████| 111/111 [00:00<00:00, 476.29it/s]\n",
      "Epoch 35/200: 100%|██████████| 111/111 [00:00<00:00, 475.73it/s]\n",
      "Epoch 36/200: 100%|██████████| 111/111 [00:00<00:00, 471.95it/s]\n",
      "Epoch 37/200: 100%|██████████| 111/111 [00:00<00:00, 476.08it/s]\n",
      "Epoch 38/200: 100%|██████████| 111/111 [00:00<00:00, 475.99it/s]\n",
      "Epoch 39/200: 100%|██████████| 111/111 [00:00<00:00, 476.61it/s]\n",
      "Epoch 40/200: 100%|██████████| 111/111 [00:00<00:00, 475.79it/s]\n",
      "Epoch 41/200: 100%|██████████| 111/111 [00:00<00:00, 475.64it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 40: Train Acc: 38.49%, Train Loss: 1.4984\n",
      "Test Acc: 25.53%, Test Loss: 1.9933\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 42/200: 100%|██████████| 111/111 [00:00<00:00, 475.93it/s]\n",
      "Epoch 43/200: 100%|██████████| 111/111 [00:00<00:00, 475.80it/s]\n",
      "Epoch 44/200: 100%|██████████| 111/111 [00:00<00:00, 475.87it/s]\n",
      "Epoch 45/200: 100%|██████████| 111/111 [00:00<00:00, 476.18it/s]\n",
      "Epoch 46/200: 100%|██████████| 111/111 [00:00<00:00, 476.30it/s]\n",
      "Epoch 47/200: 100%|██████████| 111/111 [00:00<00:00, 475.83it/s]\n",
      "Epoch 48/200: 100%|██████████| 111/111 [00:00<00:00, 475.62it/s]\n",
      "Epoch 49/200: 100%|██████████| 111/111 [00:00<00:00, 476.08it/s]\n",
      "Epoch 50/200: 100%|██████████| 111/111 [00:00<00:00, 476.37it/s]\n",
      "Epoch 51/200: 100%|██████████| 111/111 [00:00<00:00, 475.80it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 50: Train Acc: 46.09%, Train Loss: 1.3019\n",
      "Test Acc: 27.98%, Test Loss: 2.0551\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 52/200: 100%|██████████| 111/111 [00:00<00:00, 475.40it/s]\n",
      "Epoch 53/200: 100%|██████████| 111/111 [00:00<00:00, 471.22it/s]\n",
      "Epoch 54/200: 100%|██████████| 111/111 [00:00<00:00, 475.88it/s]\n",
      "Epoch 55/200: 100%|██████████| 111/111 [00:00<00:00, 475.97it/s]\n",
      "Epoch 56/200: 100%|██████████| 111/111 [00:00<00:00, 476.54it/s]\n",
      "Epoch 57/200: 100%|██████████| 111/111 [00:00<00:00, 432.53it/s]\n",
      "Epoch 58/200: 100%|██████████| 111/111 [00:00<00:00, 476.37it/s]\n",
      "Epoch 59/200: 100%|██████████| 111/111 [00:00<00:00, 476.18it/s]\n",
      "Epoch 60/200: 100%|██████████| 111/111 [00:00<00:00, 473.16it/s]\n",
      "Epoch 61/200: 100%|██████████| 111/111 [00:00<00:00, 476.04it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 60: Train Acc: 51.84%, Train Loss: 1.1824\n",
      "Test Acc: 32.14%, Test Loss: 2.0477\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 62/200: 100%|██████████| 111/111 [00:00<00:00, 475.63it/s]\n",
      "Epoch 63/200: 100%|██████████| 111/111 [00:00<00:00, 475.42it/s]\n",
      "Epoch 64/200: 100%|██████████| 111/111 [00:00<00:00, 476.50it/s]\n",
      "Epoch 65/200: 100%|██████████| 111/111 [00:00<00:00, 472.69it/s]\n",
      "Epoch 66/200: 100%|██████████| 111/111 [00:00<00:00, 475.91it/s]\n",
      "Epoch 67/200: 100%|██████████| 111/111 [00:00<00:00, 476.24it/s]\n",
      "Epoch 68/200: 100%|██████████| 111/111 [00:00<00:00, 476.19it/s]\n",
      "Epoch 69/200: 100%|██████████| 111/111 [00:00<00:00, 471.79it/s]\n",
      "Epoch 70/200: 100%|██████████| 111/111 [00:00<00:00, 477.02it/s]\n",
      "Epoch 71/200: 100%|██████████| 111/111 [00:00<00:00, 476.54it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 70: Train Acc: 55.95%, Train Loss: 1.0674\n",
      "Test Acc: 32.14%, Test Loss: 2.1111\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 72/200: 100%|██████████| 111/111 [00:00<00:00, 476.58it/s]\n",
      "Epoch 73/200: 100%|██████████| 111/111 [00:00<00:00, 476.56it/s]\n",
      "Epoch 74/200: 100%|██████████| 111/111 [00:00<00:00, 475.34it/s]\n",
      "Epoch 75/200: 100%|██████████| 111/111 [00:00<00:00, 476.46it/s]\n",
      "Epoch 76/200: 100%|██████████| 111/111 [00:00<00:00, 474.63it/s]\n",
      "Epoch 77/200: 100%|██████████| 111/111 [00:00<00:00, 477.07it/s]\n",
      "Epoch 78/200: 100%|██████████| 111/111 [00:00<00:00, 476.86it/s]\n",
      "Epoch 79/200: 100%|██████████| 111/111 [00:00<00:00, 476.95it/s]\n",
      "Epoch 80/200: 100%|██████████| 111/111 [00:00<00:00, 476.04it/s]\n",
      "Epoch 81/200: 100%|██████████| 111/111 [00:00<00:00, 476.19it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 80: Train Acc: 59.92%, Train Loss: 0.9770\n",
      "Test Acc: 34.13%, Test Loss: 2.0908\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 82/200: 100%|██████████| 111/111 [00:00<00:00, 476.47it/s]\n",
      "Epoch 83/200: 100%|██████████| 111/111 [00:00<00:00, 476.23it/s]\n",
      "Epoch 84/200: 100%|██████████| 111/111 [00:00<00:00, 476.36it/s]\n",
      "Epoch 85/200: 100%|██████████| 111/111 [00:00<00:00, 471.35it/s]\n",
      "Epoch 86/200: 100%|██████████| 111/111 [00:00<00:00, 476.84it/s]\n",
      "Epoch 87/200: 100%|██████████| 111/111 [00:00<00:00, 476.98it/s]\n",
      "Epoch 88/200: 100%|██████████| 111/111 [00:00<00:00, 476.76it/s]\n",
      "Epoch 89/200: 100%|██████████| 111/111 [00:00<00:00, 476.25it/s]\n",
      "Epoch 90/200: 100%|██████████| 111/111 [00:00<00:00, 476.19it/s]\n",
      "Epoch 91/200: 100%|██████████| 111/111 [00:00<00:00, 476.53it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 90: Train Acc: 64.03%, Train Loss: 0.8874\n",
      "Test Acc: 35.98%, Test Loss: 2.1436\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 92/200: 100%|██████████| 111/111 [00:00<00:00, 476.71it/s]\n",
      "Epoch 93/200: 100%|██████████| 111/111 [00:00<00:00, 471.71it/s]\n",
      "Epoch 94/200: 100%|██████████| 111/111 [00:00<00:00, 431.71it/s]\n",
      "Epoch 95/200: 100%|██████████| 111/111 [00:00<00:00, 476.83it/s]\n",
      "Epoch 96/200: 100%|██████████| 111/111 [00:00<00:00, 476.21it/s]\n",
      "Epoch 97/200: 100%|██████████| 111/111 [00:00<00:00, 476.77it/s]\n",
      "Epoch 98/200: 100%|██████████| 111/111 [00:00<00:00, 476.41it/s]\n",
      "Epoch 99/200: 100%|██████████| 111/111 [00:00<00:00, 476.68it/s]\n",
      "Epoch 100/200: 100%|██████████| 111/111 [00:00<00:00, 476.37it/s]\n",
      "Epoch 101/200: 100%|██████████| 111/111 [00:00<00:00, 471.85it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 100: Train Acc: 68.31%, Train Loss: 0.8010\n",
      "Test Acc: 35.52%, Test Loss: 2.2599\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 102/200: 100%|██████████| 111/111 [00:00<00:00, 474.09it/s]\n",
      "Epoch 103/200: 100%|██████████| 111/111 [00:00<00:00, 476.18it/s]\n",
      "Epoch 104/200: 100%|██████████| 111/111 [00:00<00:00, 476.35it/s]\n",
      "Epoch 105/200: 100%|██████████| 111/111 [00:00<00:00, 476.30it/s]\n",
      "Epoch 106/200: 100%|██████████| 111/111 [00:00<00:00, 475.67it/s]\n",
      "Epoch 107/200: 100%|██████████| 111/111 [00:00<00:00, 476.68it/s]\n",
      "Epoch 108/200: 100%|██████████| 111/111 [00:00<00:00, 476.54it/s]\n",
      "Epoch 109/200: 100%|██████████| 111/111 [00:00<00:00, 471.99it/s]\n",
      "Epoch 110/200: 100%|██████████| 111/111 [00:00<00:00, 476.96it/s]\n",
      "Epoch 111/200: 100%|██████████| 111/111 [00:00<00:00, 477.21it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 110: Train Acc: 72.22%, Train Loss: 0.7324\n",
      "Test Acc: 40.54%, Test Loss: 2.0601\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 112/200: 100%|██████████| 111/111 [00:00<00:00, 476.77it/s]\n",
      "Epoch 113/200: 100%|██████████| 111/111 [00:00<00:00, 476.84it/s]\n",
      "Epoch 114/200: 100%|██████████| 111/111 [00:00<00:00, 476.69it/s]\n",
      "Epoch 115/200: 100%|██████████| 111/111 [00:00<00:00, 476.74it/s]\n",
      "Epoch 116/200: 100%|██████████| 111/111 [00:00<00:00, 476.70it/s]\n",
      "Epoch 117/200: 100%|██████████| 111/111 [00:00<00:00, 472.05it/s]\n",
      "Epoch 118/200: 100%|██████████| 111/111 [00:00<00:00, 476.62it/s]\n",
      "Epoch 119/200: 100%|██████████| 111/111 [00:00<00:00, 478.15it/s]\n",
      "Epoch 120/200: 100%|██████████| 111/111 [00:00<00:00, 477.44it/s]\n",
      "Epoch 121/200: 100%|██████████| 111/111 [00:00<00:00, 476.97it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 120: Train Acc: 75.26%, Train Loss: 0.6677\n",
      "Test Acc: 42.59%, Test Loss: 2.0922\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 122/200: 100%|██████████| 111/111 [00:00<00:00, 476.69it/s]\n",
      "Epoch 123/200: 100%|██████████| 111/111 [00:00<00:00, 476.71it/s]\n",
      "Epoch 124/200: 100%|██████████| 111/111 [00:00<00:00, 476.94it/s]\n",
      "Epoch 125/200: 100%|██████████| 111/111 [00:00<00:00, 473.74it/s]\n",
      "Epoch 126/200: 100%|██████████| 111/111 [00:00<00:00, 476.88it/s]\n",
      "Epoch 127/200: 100%|██████████| 111/111 [00:00<00:00, 476.40it/s]\n",
      "Epoch 128/200: 100%|██████████| 111/111 [00:00<00:00, 475.45it/s]\n",
      "Epoch 129/200: 100%|██████████| 111/111 [00:00<00:00, 477.37it/s]\n",
      "Epoch 130/200: 100%|██████████| 111/111 [00:00<00:00, 472.73it/s]\n",
      "Epoch 131/200: 100%|██████████| 111/111 [00:00<00:00, 471.68it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 130: Train Acc: 78.15%, Train Loss: 0.6015\n",
      "Test Acc: 42.72%, Test Loss: 2.1730\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 132/200: 100%|██████████| 111/111 [00:00<00:00, 475.80it/s]\n",
      "Epoch 133/200: 100%|██████████| 111/111 [00:00<00:00, 471.92it/s]\n",
      "Epoch 134/200: 100%|██████████| 111/111 [00:00<00:00, 476.67it/s]\n",
      "Epoch 135/200: 100%|██████████| 111/111 [00:00<00:00, 475.78it/s]\n",
      "Epoch 136/200: 100%|██████████| 111/111 [00:00<00:00, 476.34it/s]\n",
      "Epoch 137/200: 100%|██████████| 111/111 [00:00<00:00, 476.60it/s]\n",
      "Epoch 138/200: 100%|██████████| 111/111 [00:00<00:00, 473.45it/s]\n",
      "Epoch 139/200: 100%|██████████| 111/111 [00:00<00:00, 475.78it/s]\n",
      "Epoch 140/200: 100%|██████████| 111/111 [00:00<00:00, 476.75it/s]\n",
      "Epoch 141/200: 100%|██████████| 111/111 [00:00<00:00, 473.41it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 140: Train Acc: 79.88%, Train Loss: 0.5724\n",
      "Test Acc: 45.70%, Test Loss: 2.0698\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 142/200: 100%|██████████| 111/111 [00:00<00:00, 476.38it/s]\n",
      "Epoch 143/200: 100%|██████████| 111/111 [00:00<00:00, 476.82it/s]\n",
      "Epoch 144/200: 100%|██████████| 111/111 [00:00<00:00, 476.67it/s]\n",
      "Epoch 145/200: 100%|██████████| 111/111 [00:00<00:00, 476.49it/s]\n",
      "Epoch 146/200: 100%|██████████| 111/111 [00:00<00:00, 476.45it/s]\n",
      "Epoch 147/200: 100%|██████████| 111/111 [00:00<00:00, 476.84it/s]\n",
      "Epoch 148/200: 100%|██████████| 111/111 [00:00<00:00, 475.30it/s]\n",
      "Epoch 149/200: 100%|██████████| 111/111 [00:00<00:00, 476.57it/s]\n",
      "Epoch 150/200: 100%|██████████| 111/111 [00:00<00:00, 471.57it/s]\n",
      "Epoch 151/200: 100%|██████████| 111/111 [00:00<00:00, 476.25it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 150: Train Acc: 81.35%, Train Loss: 0.4940\n",
      "Test Acc: 46.83%, Test Loss: 2.1886\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 152/200: 100%|██████████| 111/111 [00:00<00:00, 476.66it/s]\n",
      "Epoch 153/200: 100%|██████████| 111/111 [00:00<00:00, 476.78it/s]\n",
      "Epoch 154/200: 100%|██████████| 111/111 [00:00<00:00, 476.23it/s]\n",
      "Epoch 155/200: 100%|██████████| 111/111 [00:00<00:00, 477.04it/s]\n",
      "Epoch 156/200: 100%|██████████| 111/111 [00:00<00:00, 476.13it/s]\n",
      "Epoch 157/200: 100%|██████████| 111/111 [00:00<00:00, 476.34it/s]\n",
      "Epoch 158/200: 100%|██████████| 111/111 [00:00<00:00, 471.00it/s]\n",
      "Epoch 159/200: 100%|██████████| 111/111 [00:00<00:00, 475.35it/s]\n",
      "Epoch 160/200: 100%|██████████| 111/111 [00:00<00:00, 476.43it/s]\n",
      "Epoch 161/200: 100%|██████████| 111/111 [00:00<00:00, 475.78it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 160: Train Acc: 82.71%, Train Loss: 0.4775\n",
      "Test Acc: 50.07%, Test Loss: 1.9922\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 162/200: 100%|██████████| 111/111 [00:00<00:00, 476.47it/s]\n",
      "Epoch 163/200: 100%|██████████| 111/111 [00:00<00:00, 476.25it/s]\n",
      "Epoch 164/200: 100%|██████████| 111/111 [00:00<00:00, 476.28it/s]\n",
      "Epoch 165/200: 100%|██████████| 111/111 [00:00<00:00, 476.17it/s]\n",
      "Epoch 166/200: 100%|██████████| 111/111 [00:00<00:00, 471.60it/s]\n",
      "Epoch 167/200: 100%|██████████| 111/111 [00:00<00:00, 437.42it/s]\n",
      "Epoch 168/200: 100%|██████████| 111/111 [00:00<00:00, 472.53it/s]\n",
      "Epoch 169/200: 100%|██████████| 111/111 [00:00<00:00, 476.30it/s]\n",
      "Epoch 170/200: 100%|██████████| 111/111 [00:00<00:00, 476.23it/s]\n",
      "Epoch 171/200: 100%|██████████| 111/111 [00:00<00:00, 475.35it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 170: Train Acc: 85.06%, Train Loss: 0.4346\n",
      "Test Acc: 47.29%, Test Loss: 2.2052\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 172/200: 100%|██████████| 111/111 [00:00<00:00, 475.92it/s]\n",
      "Epoch 173/200: 100%|██████████| 111/111 [00:00<00:00, 475.78it/s]\n",
      "Epoch 174/200: 100%|██████████| 111/111 [00:00<00:00, 471.42it/s]\n",
      "Epoch 175/200: 100%|██████████| 111/111 [00:00<00:00, 473.15it/s]\n",
      "Epoch 176/200: 100%|██████████| 111/111 [00:00<00:00, 475.65it/s]\n",
      "Epoch 177/200: 100%|██████████| 111/111 [00:00<00:00, 476.09it/s]\n",
      "Epoch 178/200: 100%|██████████| 111/111 [00:00<00:00, 476.60it/s]\n",
      "Epoch 179/200: 100%|██████████| 111/111 [00:00<00:00, 476.28it/s]\n",
      "Epoch 180/200: 100%|██████████| 111/111 [00:00<00:00, 476.24it/s]\n",
      "Epoch 181/200: 100%|██████████| 111/111 [00:00<00:00, 475.82it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 180: Train Acc: 82.79%, Train Loss: 0.4665\n",
      "Test Acc: 51.79%, Test Loss: 2.1221\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 182/200: 100%|██████████| 111/111 [00:00<00:00, 471.39it/s]\n",
      "Epoch 183/200: 100%|██████████| 111/111 [00:00<00:00, 476.23it/s]\n",
      "Epoch 184/200: 100%|██████████| 111/111 [00:00<00:00, 476.07it/s]\n",
      "Epoch 185/200: 100%|██████████| 111/111 [00:00<00:00, 476.21it/s]\n",
      "Epoch 186/200: 100%|██████████| 111/111 [00:00<00:00, 476.20it/s]\n",
      "Epoch 187/200: 100%|██████████| 111/111 [00:00<00:00, 476.05it/s]\n",
      "Epoch 188/200: 100%|██████████| 111/111 [00:00<00:00, 475.85it/s]\n",
      "Epoch 189/200: 100%|██████████| 111/111 [00:00<00:00, 476.09it/s]\n",
      "Epoch 190/200: 100%|██████████| 111/111 [00:00<00:00, 471.55it/s]\n",
      "Epoch 191/200: 100%|██████████| 111/111 [00:00<00:00, 475.72it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 190: Train Acc: 85.63%, Train Loss: 0.4117\n",
      "Test Acc: 50.79%, Test Loss: 2.1503\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 192/200: 100%|██████████| 111/111 [00:00<00:00, 475.59it/s]\n",
      "Epoch 193/200: 100%|██████████| 111/111 [00:00<00:00, 474.93it/s]\n",
      "Epoch 194/200: 100%|██████████| 111/111 [00:00<00:00, 476.34it/s]\n",
      "Epoch 195/200: 100%|██████████| 111/111 [00:00<00:00, 476.07it/s]\n",
      "Epoch 196/200: 100%|██████████| 111/111 [00:00<00:00, 475.64it/s]\n",
      "Epoch 197/200: 100%|██████████| 111/111 [00:00<00:00, 475.78it/s]\n",
      "Epoch 198/200: 100%|██████████| 111/111 [00:00<00:00, 470.96it/s]\n",
      "Epoch 199/200: 100%|██████████| 111/111 [00:00<00:00, 475.77it/s]\n",
      "Epoch 200/200: 100%|██████████| 111/111 [00:00<00:00, 475.42it/s]\n"
     ]
    }
   ],
   "source": [
    "print(f'Training the MLP for {representation}, {statistic}...')\n",
    "best_accuracy = 0\n",
    "\n",
    "# Create and train the model with the specified hyperparameters\n",
    "model = FeedForward(dimensions=[input_size, 128, 128, 64], num_classes=output_size, dropout_rate=DROPOUT)\n",
    "model.to(device)\n",
    "criterion = nn.CrossEntropyLoss()\n",
    "optimizer = torch.optim.Adam(model.parameters(), lr=LR, weight_decay=WEIGHT_DECAY)\n",
    "\n",
    "trainloader = DataLoader(trainset, batch_size=BATCH_SIZE, shuffle=True)\n",
    "testloader = DataLoader(testset, batch_size=BATCH_SIZE, shuffle=False)\n",
    "\n",
    "# Train the model\n",
    "all_train_loss, all_train_acc, all_test_loss, all_test_acc = train( model, trainloader, testloader, criterion, \n",
    "                                                                   optimizer, epochs=EPOCHS, bs=BATCH_SIZE, when_to_print=10)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 107,
   "id": "2c9efd04-e818-4805-8113-3124fa0b0550",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 800x550 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.title(f'Loss curve for representation:{representation},\\n statistic:{statistic},\\n N = {N}')\n",
    "plt.xlabel('EPOCH')\n",
    "plt.ylabel('Loss')\n",
    "plt.plot(all_train_loss, label = 'train data')\n",
    "plt.plot(all_test_loss, label= 'test data')\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 108,
   "id": "d2daff42-740a-43f0-9c41-e6eb525b1256",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 800x550 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.title(f'Model accuracy for representation:{representation},\\n statistic:{statistic},\\n N = {N}')\n",
    "plt.xlabel('EPOCH')\n",
    "plt.ylabel('Accuracy')\n",
    "plt.plot(all_train_acc, label = 'train data')\n",
    "plt.plot(all_test_acc, label= 'test data')\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "8e3041df-42c0-4e08-b298-db80477bf2fb",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "permutation",
   "language": "python",
   "name": "permutation"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.11.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}