{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "d2a402ac-ec71-4cf5-b802-bd758a603ef7",
   "metadata": {},
   "outputs": [],
   "source": [
    "from rbm.params import EntityParams\n",
    "from rbm.layer import Layer\n",
    "from rbm.status import Status\n",
    "from rbm.train import train\n",
    "from rbm.matrix import Mat, np"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "976e0d6b-34b3-4745-92df-14416e043661",
   "metadata": {},
   "outputs": [],
   "source": [
    "params = EntityParams()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "c926b611-bb14-47b9-b870-b95cb5480fca",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'learning_rate': 0.1,\n",
       " 'momentum': 0.5,\n",
       " 'weight_decay': 0,\n",
       " 'num_epochs': 1000,\n",
       " 'mini_batch_size': 0,\n",
       " 'do_rao_blackwell': False,\n",
       " 'do_gibbs_sample_visible': False,\n",
       " 'do_gibbs_sample_hidden': False,\n",
       " 'do_batch_sample': False,\n",
       " 'num_gibbs_samples': 1,\n",
       " 'do_gaussian_visible': False,\n",
       " 'do_gaussian_hidden': False}"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "params.__dict__"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "1ad7ae45-6daf-4f89-a5b4-b6b790c4f76c",
   "metadata": {},
   "outputs": [],
   "source": [
    "params.do_rao_blackwell = True\n",
    "params.num_gibbs_samples = 3"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "f7ca5d0c-687a-40f9-942a-012583efe7c8",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'learning_rate': 0.1,\n",
       " 'momentum': 0.5,\n",
       " 'weight_decay': 0,\n",
       " 'num_epochs': 1000,\n",
       " 'mini_batch_size': 0,\n",
       " 'do_rao_blackwell': True,\n",
       " 'do_gibbs_sample_visible': False,\n",
       " 'do_gibbs_sample_hidden': False,\n",
       " 'do_batch_sample': False,\n",
       " 'num_gibbs_samples': 3,\n",
       " 'do_gaussian_visible': False,\n",
       " 'do_gaussian_hidden': False}"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "params.__dict__"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "30aff23e-6603-426b-aeb6-000fc70e109f",
   "metadata": {},
   "outputs": [],
   "source": [
    "layer = Layer(\"Layer_0\", (3, 1, 0, 16), params)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "510322fa-c74d-4493-b81c-a4c9cb650078",
   "metadata": {},
   "outputs": [],
   "source": [
    "layer.init(0.01)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "aa1bdd0e-a298-46c2-8bff-1f614be8fa51",
   "metadata": {},
   "outputs": [],
   "source": [
    "layer.load()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "dc92799d-2b53-402a-a51d-285c4f51126a",
   "metadata": {},
   "outputs": [],
   "source": [
    "training_batch = Mat([[0,1,1], [0,0,0], [1,1,0], [1,0,1]], dtype=np.float64)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "1d027847-7362-46b5-b75a-b9de09f10602",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "-------------------------------------------\n",
      "progress              : 0%\n",
      "err_rms              : 0.24999708676779586\n",
      "-------------------------------------------\n",
      "progress              : 10%\n",
      "err_rms              : 0.24995450907424\n",
      "-------------------------------------------\n",
      "progress              : 20%\n",
      "err_rms              : 0.24948447203582114\n",
      "-------------------------------------------\n",
      "progress              : 30%\n",
      "err_rms              : 0.24412963558981704\n",
      "-------------------------------------------\n",
      "progress              : 40%\n",
      "err_rms              : 0.1949812378035757\n",
      "-------------------------------------------\n",
      "progress              : 50%\n",
      "err_rms              : 0.048675091786564394\n",
      "-------------------------------------------\n",
      "progress              : 60%\n",
      "err_rms              : 0.00301183842639605\n",
      "-------------------------------------------\n",
      "progress              : 70%\n",
      "err_rms              : 0.0006911957660915791\n",
      "-------------------------------------------\n",
      "progress              : 80%\n",
      "err_rms              : 0.000293184530139503\n",
      "-------------------------------------------\n",
      "progress              : 90%\n",
      "err_rms              : 0.0001607680838909887\n",
      "-------------------------------------------\n",
      "progress              : 100%\n",
      "err_rms              : 0.00010079934874358073\n",
      "-------------------------------------------\n",
      "progress              : 100%\n",
      "err_rms_total              : 9.954453566455675e-05\n"
     ]
    }
   ],
   "source": [
    "train(layer.entity, training_batch, Status())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "fb4c4488-2137-440c-98b6-d07e73356a0a",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "P[0. 0. 0.] : [[0.01017997 0.01309764 0.01289942]]\n",
      "P[0. 1. 0.] : [[0.01382953 0.99673111 0.0121298 ]]\n",
      "P[1. 0. 0.] : [[0.99411057 0.03544156 0.02632994]]\n",
      "P[1. 1. 0.] : [[0.99256076 0.98906099 0.00948805]]\n"
     ]
    }
   ],
   "source": [
    "# Test with test data\n",
    "test_batch = Mat([[0,0,0], [0,1,0], [1,0,0], [1,1,0]], dtype=np.float64)\n",
    "for pattern in test_batch:\n",
    "\th = layer.entity.forward(pattern)\n",
    "\tv = layer.entity.reconstruct(h)\n",
    "\tprint(f\"P{pattern} : {v}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "a9502c36-894d-4b33-90f0-a5a3ea9fcc4e",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[0., 0., 0.],\n",
       "       [0., 1., 0.],\n",
       "       [1., 0., 0.],\n",
       "       [1., 1., 0.]])"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_batch"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "39dcac95-a132-40f9-9dfc-cb8161e7f881",
   "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": "3.12.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}