circomlib-ml/models/conv1d.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "from tensorflow.keras.layers import Input, Conv1D\n",
    "from tensorflow.keras import Model\n",
    "import numpy as np"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "inputs = Input(shape=(20,3))\n",
    "x = Conv1D(2,4,3)(inputs)\n",
    "model = Model(inputs, x)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"model\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "input_1 (InputLayer)         [(None, 20, 3)]           0         \n",
      "_________________________________________________________________\n",
      "conv1d (Conv1D)              (None, 6, 2)              26        \n",
      "=================================================================\n",
      "Total params: 26\n",
      "Trainable params: 26\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<tf.Variable 'conv1d/kernel:0' shape=(4, 3, 2) dtype=float32, numpy=\n",
       " array([[[ 0.13351655,  0.15176392],\n",
       "         [ 0.02279764, -0.05288953],\n",
       "         [ 0.31283128, -0.2973013 ]],\n",
       " \n",
       "        [[ 0.27222842, -0.14998454],\n",
       "         [ 0.10822219, -0.36693448],\n",
       "         [ 0.14209783, -0.29082325]],\n",
       " \n",
       "        [[-0.48623034, -0.4235212 ],\n",
       "         [ 0.07678127,  0.32315302],\n",
       "         [-0.11050957, -0.03980196]],\n",
       " \n",
       "        [[ 0.26589322,  0.00770217],\n",
       "         [ 0.11929381,  0.29532135],\n",
       "         [-0.42807332,  0.03231919]]], dtype=float32)>,\n",
       " <tf.Variable 'conv1d/bias:0' shape=(2,) dtype=float32, numpy=array([0., 0.], dtype=float32)>]"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.weights"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[[0.84452152, 0.40819381, 0.90425158],\n",
       "        [0.65092274, 0.59893334, 0.60028247],\n",
       "        [0.44949689, 0.254961  , 0.40024966],\n",
       "        [0.57719296, 0.96000249, 0.08217882],\n",
       "        [0.66562102, 0.00446413, 0.6464117 ],\n",
       "        [0.85546508, 0.38582714, 0.8505983 ],\n",
       "        [0.27450673, 0.201367  , 0.18818527],\n",
       "        [0.90095107, 0.33781381, 0.84773899],\n",
       "        [0.36179829, 0.39354172, 0.64100907],\n",
       "        [0.35045615, 0.37234526, 0.48415795],\n",
       "        [0.11139122, 0.60499841, 0.58387442],\n",
       "        [0.87088195, 0.64640523, 0.48402816],\n",
       "        [0.9638806 , 0.29440207, 0.22027009],\n",
       "        [0.3764113 , 0.13575624, 0.7720717 ],\n",
       "        [0.42784105, 0.41073501, 0.28311926],\n",
       "        [0.66622245, 0.95378854, 0.66375939],\n",
       "        [0.41543282, 0.37529526, 0.23072244],\n",
       "        [0.54334445, 0.1388458 , 0.85307472],\n",
       "        [0.25258015, 0.37725648, 0.75018739],\n",
       "        [0.75532678, 0.52800654, 0.46152891]]])"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X = np.random.rand(1,20,3)\n",
    "X"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[[ 0.7218593 , -0.4875301 ],\n",
       "        [-0.06567734, -0.48088893],\n",
       "        [ 0.21620643, -0.45382428],\n",
       "        [ 0.1553162 , -0.5966817 ],\n",
       "        [ 0.23019192, -0.01706157],\n",
       "        [-0.05194061, -0.49539083]]], dtype=float32)"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y = model.predict(X)\n",
    "y"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "in_json = {\n",
    "    \"in\": (X*1000).round().astype(int).flatten().tolist(),\n",
    "    \"weights\": (model.weights[0].numpy()*1000).round().astype(int).flatten().tolist(),\n",
    "    \"bias\": np.zeros(model.weights[1].numpy().shape).tolist()\n",
    "}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "out_json = {\n",
    "    \"out\": (y*1000000).round().astype(int).flatten().tolist()\n",
    "}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "import json"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open(\"conv1D_input.json\", \"w\") as f:\n",
    "    json.dump(in_json, f)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open(\"conv1D_output.json\", \"w\") as f:\n",
    "    json.dump(out_json, f)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "tf24",
   "language": "python",
   "name": "tf24"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.6"
  },
  "orig_nbformat": 4
 },
 "nbformat": 4,
 "nbformat_minor": 2
}