Demo I Sec 4.3 NNBits on the Speck32 Avalanche Dataset.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "8f70d02d",
   "metadata": {
    "pycharm": {
     "name": "#%% md\n"
    }
   },
   "source": [
    "# Generate Avalanche Data "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "49986660",
   "metadata": {
    "pycharm": {
     "name": "#%% md\n"
    }
   },
   "source": [
    "Create a folder for the data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "4ddea48f",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-05-19T08:15:56.904775Z",
     "start_time": "2022-05-19T08:15:56.688420Z"
    },
    "execution": {
     "iopub.execute_input": "2023-02-10T06:40:40.268894Z",
     "iopub.status.busy": "2023-02-10T06:40:40.268650Z",
     "iopub.status.idle": "2023-02-10T06:40:40.485976Z",
     "shell.execute_reply": "2023-02-10T06:40:40.485474Z",
     "shell.execute_reply.started": "2023-02-10T06:40:40.268833Z"
    },
    "pycharm": {
     "name": "#%%\n"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "mkdir: cannot create directory ‘speck_32_64’: File exists\n"
     ]
    }
   ],
   "source": [
    "!mkdir 'speck_32_64'"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e0192030",
   "metadata": {
    "pycharm": {
     "name": "#%% md\n"
    }
   },
   "source": [
    "Generate `number_of_samples = 300_000` avalanche units (this takes about one minute):"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "ede1541f",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-05-19T08:18:02.564090Z",
     "start_time": "2022-05-19T08:16:33.289946Z"
    },
    "execution": {
     "iopub.execute_input": "2023-02-10T06:40:41.284467Z",
     "iopub.status.busy": "2023-02-10T06:40:41.284204Z",
     "iopub.status.idle": "2023-02-10T06:41:19.329530Z",
     "shell.execute_reply": "2023-02-10T06:41:19.328381Z",
     "shell.execute_reply.started": "2023-02-10T06:40:41.284448Z"
    },
    "pycharm": {
     "name": "#%%\n"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "number_of_samples = 300_000\n",
    "\n",
    "from avalanche_data_generator.speck_32_64 import speck_k64_p32_o32_r22 as data_generator\n",
    "import numpy as np\n",
    "\n",
    "dataset = data_generator.generate_avalanche_dataset(int(number_of_samples))\n",
    "\n",
    "np.save(f\"speck_32_64/round6_sequences300k.npy\", dataset[6]) # please note that we start counting the rounds from 0, i.e. round id 6 means analyzing round 7/22 "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "77617abc-04c3-4664-a8f3-741b7b2ec66a",
   "metadata": {},
   "source": [
    "# Choose Workspace Folder"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "63bf4057-7eee-47cb-a7c7-7664454c1dc1",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-02-10T06:41:19.332029Z",
     "iopub.status.busy": "2023-02-10T06:41:19.331529Z",
     "iopub.status.idle": "2023-02-10T06:41:19.335162Z",
     "shell.execute_reply": "2023-02-10T06:41:19.334783Z",
     "shell.execute_reply.started": "2023-02-10T06:41:19.332006Z"
    }
   },
   "outputs": [],
   "source": [
    "# choose where to save the analysis data \n",
    "# this folder contains the configuration file, which points out the dataset and GPU configuration to use\n",
    "savepath = 'demo_speck32_round7'"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c66df706",
   "metadata": {
    "pycharm": {
     "name": "#%% md\n"
    }
   },
   "source": [
    "# Check Hardware Resources\n",
    "The available resources can be explored by using for example the `ray.available_resources()` or `nvidia-smi`:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "d89ebfd7",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-06-04T06:54:45.203333Z",
     "start_time": "2022-06-04T06:54:41.596826Z"
    },
    "execution": {
     "iopub.execute_input": "2023-02-10T06:41:19.336102Z",
     "iopub.status.busy": "2023-02-10T06:41:19.335986Z",
     "iopub.status.idle": "2023-02-10T06:41:24.342405Z",
     "shell.execute_reply": "2023-02-10T06:41:24.342001Z",
     "shell.execute_reply.started": "2023-02-10T06:41:19.336087Z"
    },
    "pycharm": {
     "name": "#%%\n"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2023-02-10 06:41:21,789\tINFO services.py:1245 -- View the Ray dashboard at \u001B[1m\u001B[32mhttp://127.0.0.1:8265\u001B[39m\u001B[22m\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "{'memory': 338659682919.0,\n",
       " 'node:172.31.102.81': 1.0,\n",
       " 'CPU': 128.0,\n",
       " 'GPU': 4.0,\n",
       " 'object_store_memory': 149425578393.0}"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import ray\n",
    "ray.init()\n",
    "ray.available_resources()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "cabc71e8",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-06-04T06:54:46.077342Z",
     "start_time": "2022-06-04T06:54:45.207831Z"
    },
    "execution": {
     "iopub.execute_input": "2023-02-10T06:41:24.343816Z",
     "iopub.status.busy": "2023-02-10T06:41:24.343617Z",
     "iopub.status.idle": "2023-02-10T06:41:25.435581Z",
     "shell.execute_reply": "2023-02-10T06:41:25.435060Z",
     "shell.execute_reply.started": "2023-02-10T06:41:24.343791Z"
    },
    "pycharm": {
     "name": "#%%\n"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fri Feb 10 06:41:24 2023       \n",
      "+-----------------------------------------------------------------------------+\n",
      "| NVIDIA-SMI 450.80.02    Driver Version: 450.80.02    CUDA Version: 11.0     |\n",
      "|-------------------------------+----------------------+----------------------+\n",
      "| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |\n",
      "| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |\n",
      "|                               |                      |               MIG M. |\n",
      "|===============================+======================+======================|\n",
      "|   0  A100-SXM4-40GB      On   | 00000000:01:00.0 Off |                    0 |\n",
      "| N/A   34C    P0    52W / 275W |   9725MiB / 40537MiB |      0%      Default |\n",
      "|                               |                      |             Disabled |\n",
      "+-------------------------------+----------------------+----------------------+\n",
      "|   1  A100-SXM4-40GB      On   | 00000000:47:00.0 Off |                    0 |\n",
      "| N/A   34C    P0    53W / 275W |      3MiB / 40537MiB |      0%      Default |\n",
      "|                               |                      |             Disabled |\n",
      "+-------------------------------+----------------------+----------------------+\n",
      "|   2  A100-SXM4-40GB      On   | 00000000:81:00.0 Off |                    0 |\n",
      "| N/A   35C    P0    56W / 275W |      3MiB / 40537MiB |      0%      Default |\n",
      "|                               |                      |             Disabled |\n",
      "+-------------------------------+----------------------+----------------------+\n",
      "|   3  A100-SXM4-40GB      On   | 00000000:C2:00.0 Off |                    0 |\n",
      "| N/A   35C    P0    55W / 275W |      3MiB / 40537MiB |      0%      Default |\n",
      "|                               |                      |             Disabled |\n",
      "+-------------------------------+----------------------+----------------------+\n",
      "                                                                               \n",
      "+-----------------------------------------------------------------------------+\n",
      "| Processes:                                                                  |\n",
      "|  GPU   GI   CI        PID   Type   Process name                  GPU Memory |\n",
      "|        ID   ID                                                   Usage      |\n",
      "|=============================================================================|\n",
      "+-----------------------------------------------------------------------------+\n"
     ]
    }
   ],
   "source": [
    "!nvidia-smi"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1c6095bd",
   "metadata": {
    "pycharm": {
     "name": "#%% md\n"
    }
   },
   "source": [
    "# Run NNBits Analysis"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "f684e2ad",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-06-05T11:47:47.713374Z",
     "start_time": "2022-06-05T11:46:53.810986Z"
    },
    "execution": {
     "iopub.execute_input": "2023-02-10T06:45:04.666321Z",
     "iopub.status.busy": "2023-02-10T06:45:04.666100Z",
     "iopub.status.idle": "2023-02-10T06:55:07.842494Z",
     "shell.execute_reply": "2023-02-10T06:55:07.841851Z",
     "shell.execute_reply.started": "2023-02-10T06:45:04.666302Z"
    },
    "pycharm": {
     "name": "#%%\n"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2023-02-10_06h45m07s \t started to load data from harddisk...\n",
      "2023-02-10_06h45m07s \t finished.\n",
      "======================================================================\n",
      "|                             | value                                |\n",
      "|:----------------------------|:-------------------------------------|\n",
      "| DATAPATH                    | speck_32_64/round6_sequences300k.npy |\n",
      "| NEURAL_NETWORK_MODEL        | gohr_generalized                     |\n",
      "| NEURAL_NETWORKS             | 100                                  |\n",
      "| N_RANDOM_BITS               | 63                                   |\n",
      "| N_GPUS                      | 1                                    |\n",
      "| N_ACTORS_PER_GPU            | 4                                    |\n",
      "| GPU_PER_ACTOR               | 0.25                                 |\n",
      "| CPU_PER_ACTOR               | 5                                    |\n",
      "| N_EPOCHS                    | 5                                    |\n",
      "| N_TRAIN                     | 145000                               |\n",
      "| N_VAL                       | 145000                               |\n",
      "| N_TEST                      | 0                                    |\n",
      "| BATCHSIZE                   | 5000                                 |\n",
      "| INPUT_DATA_OP               | zero                                 |\n",
      "| SELECT_BITS_STRATEGY        | random                               |\n",
      "| PREDICT_LABEL               | False                                |\n",
      "| TARGET_BITS                 | []                                   |\n",
      "| EARLY_STOPPING              | True                                 |\n",
      "| EARLY_STOPPING_P_VALUE      | 1e-10                                |\n",
      "| CREATE_NEW_SELECTIONS       | True                                 |\n",
      "| MODEL_STRENGTH              | 1                                    |\n",
      "| SAVE_WEIGHTS                | False                                |\n",
      "| SAVE_BEST_WEIGHTS           | False                                |\n",
      "| TESTING_BATCH_SIZE          | 0                                    |\n",
      "| TEST_ONLY                   | False                                |\n",
      "| RESULTING BITS IN DATA-ROW  | 1024                                 |\n",
      "| RESULTING N SELECTED BITS   | 63                                   |\n",
      "| RESULTING N TOTAL BITS      | 1024                                 |\n",
      "| RESULTING NN INPUT NEURONS  | 1024                                 |\n",
      "| RESULTING NN OUTPUT NEURONS | 63                                   |\n",
      "| VALIDATION_BATCH_SIZE       |                                      |\n",
      "\n",
      "\n",
      "2023-02-10 06:45:09,107\tINFO services.py:1245 -- View the Ray dashboard at \u001B[1m\u001B[32mhttp://127.0.0.1:8266\u001B[39m\u001B[22m\n",
      "||           time            |   NN finished   | pred. bits ||  best bit  |  acc (%)   |   n pred   |  p value   ||\n",
      "===================================================================================================================\n",
      "||   2023-02-10_06h45m11s    |      0/100      |   0/1024   ||    nan     |    nan     |    nan     |    nan     ||\n",
      "||   2023-02-10_06h45m42s    |      1/100      |  63/1024   ||    715     |   50.255   |     1      |   0.052    ||\n",
      "||   2023-02-10_06h45m42s    |      2/100      |  126/1024  ||    440     |   50.317   |     1      |   0.016    ||\n",
      "||   2023-02-10_06h45m43s    |      3/100      |  189/1024  ||    820     |   50.321   |     1      |   0.015    ||\n",
      "||   2023-02-10_06h45m43s    |      4/100      |  252/1024  ||    716     |   50.821   |     1      |   4e-10    ||\n",
      "||   2023-02-10_06h46m03s    |      5/100      |  315/1024  ||    716     |   50.821   |     1      |   4e-10    ||\n",
      "||   2023-02-10_06h46m06s    |      6/100      |  377/1024  ||    716     |   50.821   |     1      |   4e-10    ||\n",
      "||   2023-02-10_06h46m06s    |      7/100      |  440/1024  ||    716     |   50.821   |     1      |   4e-10    ||\n",
      "||   2023-02-10_06h46m07s    |      8/100      |  503/1024  ||    716     |   50.821   |     1      |   4e-10    ||\n",
      "||   2023-02-10_06h46m25s    |      9/100      |  549/1024  ||    716     |   50.821   |     1      |   4e-10    ||\n",
      "||   2023-02-10_06h46m27s    |     10/100      |  589/1024  ||    716     |   50.821   |     1      |   4e-10    ||\n",
      "||   2023-02-10_06h46m28s    |     11/100      |  627/1024  ||    716     |   50.646   |     2      |  8.7e-07   ||\n",
      "||   2023-02-10_06h46m28s    |     12/100      |  665/1024  ||    716     |   50.646   |     2      |  8.7e-07   ||\n",
      "||   2023-02-10_06h46m48s    |     13/100      |  697/1024  ||    716     |   50.646   |     2      |  8.7e-07   ||\n",
      "||   2023-02-10_06h46m49s    |     14/100      |  731/1024  ||    716     |   50.646   |     2      |  8.7e-07   ||\n",
      "||   2023-02-10_06h46m50s    |     15/100      |  764/1024  ||    716     |   50.646   |     2      |  8.7e-07   ||\n",
      "||   2023-02-10_06h46m50s    |     16/100      |  800/1024  ||    716     |   50.646   |     2      |  8.7e-07   ||\n",
      "||   2023-02-10_06h47m11s    |     17/100      |  826/1024  ||    716     |   50.646   |     2      |  8.7e-07   ||\n",
      "||   2023-02-10_06h47m11s    |     18/100      |  846/1024  ||    716     |   50.646   |     2      |  8.7e-07   ||\n",
      "||   2023-02-10_06h47m12s    |     20/100      |  870/1024  ||    716     |   50.646   |     2      |  8.7e-07   ||\n",
      "||   2023-02-10_06h47m12s    |     20/100      |  890/1024  ||    716     |   50.646   |     2      |  8.7e-07   ||\n",
      "||   2023-02-10_06h47m34s    |     21/100      |  907/1024  ||    716     |   50.646   |     2      |  8.7e-07   ||\n",
      "||   2023-02-10_06h47m34s    |     22/100      |  929/1024  ||    716     |   50.646   |     2      |  8.7e-07   ||\n",
      "||   2023-02-10_06h47m34s    |     23/100      |  936/1024  ||    716     |   50.646   |     2      |  8.7e-07   ||\n",
      "||   2023-02-10_06h47m34s    |     24/100      |  943/1024  ||    716     |   50.646   |     2      |  8.7e-07   ||\n",
      "||   2023-02-10_06h47m57s    |     25/100      |  952/1024  ||    716     |   50.646   |     2      |  8.7e-07   ||\n",
      "||   2023-02-10_06h47m57s    |     26/100      |  958/1024  ||    716     |   50.646   |     2      |  8.7e-07   ||\n",
      "||   2023-02-10_06h47m58s    |     27/100      |  969/1024  ||    716     |   50.646   |     2      |  8.7e-07   ||\n",
      "||   2023-02-10_06h47m58s    |     28/100      |  974/1024  ||    716     |   50.646   |     2      |  8.7e-07   ||\n",
      "||   2023-02-10_06h48m19s    |     29/100      |  979/1024  ||    716     |   50.646   |     2      |  8.7e-07   ||\n",
      "||   2023-02-10_06h48m20s    |     30/100      |  985/1024  ||    716     |   50.646   |     2      |  8.7e-07   ||\n",
      "||   2023-02-10_06h48m20s    |     32/100      |  992/1024  ||    716     |   50.646   |     2      |  8.7e-07   ||\n",
      "||   2023-02-10_06h48m20s    |     32/100      | 1000/1024  ||    716     |   50.646   |     2      |  8.7e-07   ||\n",
      "||   2023-02-10_06h48m43s    |     35/100      | 1005/1024  ||    716     |   50.646   |     2      |  8.7e-07   ||\n",
      "||   2023-02-10_06h48m43s    |     35/100      | 1009/1024  ||    716     |   50.646   |     2      |  8.7e-07   ||\n",
      "||   2023-02-10_06h48m43s    |     36/100      | 1014/1024  ||    716     |   50.646   |     2      |  8.7e-07   ||\n",
      "||   2023-02-10_06h49m08s    |     38/100      | 1017/1024  ||    716     |   50.646   |     2      |  8.7e-07   ||\n",
      "||   2023-02-10_06h49m08s    |     39/100      | 1018/1024  ||    716     |   50.646   |     2      |  8.7e-07   ||\n",
      "||   2023-02-10_06h49m08s    |     40/100      | 1020/1024  ||    716     |   50.646   |     2      |  8.7e-07   ||\n",
      "||   2023-02-10_06h49m31s    |     41/100      | 1021/1024  ||    716     |   50.646   |     2      |  8.7e-07   ||\n",
      "||   2023-02-10_06h49m32s    |     44/100      | 1022/1024  ||    716     |   50.633   |     3      |  1.4e-06   ||\n",
      "||   2023-02-10_06h49m32s    |     44/100      | 1023/1024  ||    716     |   50.633   |     3      |  1.4e-06   ||\n",
      "||   2023-02-10_06h49m55s    |     45/100      | 1023/1024  ||    716     |   50.599   |     4      |  5.2e-06   ||\n",
      "||   2023-02-10_06h49m55s    |     46/100      | 1023/1024  ||    716     |   50.599   |     4      |  5.2e-06   ||\n",
      "||   2023-02-10_06h49m56s    |     47/100      | 1024/1024  ||    716     |   50.599   |     4      |  5.2e-06   ||\n",
      "||   2023-02-10_06h49m56s    |     48/100      | 1024/1024  ||    716     |   50.599   |     4      |  5.2e-06   ||\n",
      "||   2023-02-10_06h50m19s    |     49/100      | 1024/1024  ||    716     |   50.599   |     4      |  5.2e-06   ||\n",
      "||   2023-02-10_06h50m19s    |     50/100      | 1024/1024  ||    716     |   50.599   |     4      |  5.2e-06   ||\n",
      "||   2023-02-10_06h50m19s    |     51/100      | 1024/1024  ||    716     |   50.599   |     4      |  5.2e-06   ||\n",
      "||   2023-02-10_06h50m19s    |     52/100      | 1024/1024  ||    716     |   50.599   |     4      |  5.2e-06   ||\n",
      "||   2023-02-10_06h50m43s    |     54/100      | 1024/1024  ||    716     |   50.599   |     4      |  5.2e-06   ||\n",
      "||   2023-02-10_06h50m43s    |     56/100      | 1024/1024  ||    716     |   50.599   |     4      |  5.2e-06   ||\n",
      "||   2023-02-10_06h51m06s    |     57/100      | 1024/1024  ||    716     |   50.599   |     4      |  5.2e-06   ||\n",
      "||   2023-02-10_06h51m06s    |     58/100      | 1024/1024  ||    716     |   50.659   |     5      |  5.2e-07   ||\n",
      "||   2023-02-10_06h51m06s    |     59/100      | 1024/1024  ||    716     |   50.659   |     5      |  5.2e-07   ||\n",
      "||   2023-02-10_06h51m07s    |     60/100      | 1024/1024  ||    716     |   50.659   |     5      |  5.2e-07   ||\n",
      "||   2023-02-10_06h51m30s    |     61/100      | 1024/1024  ||    716     |   50.659   |     5      |  5.2e-07   ||\n",
      "||   2023-02-10_06h51m31s    |     62/100      | 1024/1024  ||    716     |   50.659   |     5      |  5.2e-07   ||\n",
      "||   2023-02-10_06h51m31s    |     64/100      | 1024/1024  ||    716     |   50.669   |     6      |  3.6e-07   ||\n",
      "||   2023-02-10_06h51m54s    |     65/100      | 1024/1024  ||    716     |   50.669   |     6      |  3.6e-07   ||\n",
      "||   2023-02-10_06h51m54s    |     66/100      | 1024/1024  ||    716     |   50.669   |     6      |  3.6e-07   ||\n",
      "||   2023-02-10_06h51m54s    |     67/100      | 1024/1024  ||    716     |   50.669   |     6      |  3.6e-07   ||\n",
      "||   2023-02-10_06h51m54s    |     68/100      | 1024/1024  ||    716     |   50.669   |     6      |  3.6e-07   ||\n",
      "||   2023-02-10_06h52m17s    |     69/100      | 1024/1024  ||    716     |   50.669   |     6      |  3.6e-07   ||\n",
      "||   2023-02-10_06h52m17s    |     70/100      | 1024/1024  ||    716     |   50.669   |     6      |  3.6e-07   ||\n",
      "||   2023-02-10_06h52m17s    |     72/100      | 1024/1024  ||    716     |   50.669   |     6      |  3.6e-07   ||\n",
      "||   2023-02-10_06h52m40s    |     73/100      | 1024/1024  ||    716     |   50.669   |     6      |  3.6e-07   ||\n",
      "||   2023-02-10_06h52m41s    |     76/100      | 1024/1024  ||    716     |   50.669   |     6      |  3.6e-07   ||\n",
      "||   2023-02-10_06h53m04s    |     77/100      | 1024/1024  ||    716     |   50.669   |     6      |  3.6e-07   ||\n",
      "||   2023-02-10_06h53m04s    |     78/100      | 1024/1024  ||    716     |   50.669   |     6      |  3.6e-07   ||\n",
      "||   2023-02-10_06h53m05s    |     79/100      | 1024/1024  ||    716     |   50.669   |     6      |  3.6e-07   ||\n",
      "||   2023-02-10_06h53m05s    |     80/100      | 1024/1024  ||    716     |   50.669   |     6      |  3.6e-07   ||\n",
      "||   2023-02-10_06h53m28s    |     81/100      | 1024/1024  ||    716     |   50.669   |     6      |  3.6e-07   ||\n",
      "||   2023-02-10_06h53m28s    |     82/100      | 1024/1024  ||    716     |   50.669   |     6      |  3.6e-07   ||\n",
      "||   2023-02-10_06h53m28s    |     83/100      | 1024/1024  ||    716     |   50.669   |     6      |  3.6e-07   ||\n",
      "||   2023-02-10_06h53m28s    |     84/100      | 1024/1024  ||    716     |   50.669   |     6      |  3.6e-07   ||\n",
      "||   2023-02-10_06h53m51s    |     85/100      | 1024/1024  ||    716     |   50.669   |     6      |  3.6e-07   ||\n",
      "||   2023-02-10_06h53m53s    |     86/100      | 1024/1024  ||    716     |   50.669   |     6      |  3.6e-07   ||\n",
      "||   2023-02-10_06h53m53s    |     87/100      | 1024/1024  ||    716     |   50.669   |     6      |  3.6e-07   ||\n",
      "||   2023-02-10_06h53m53s    |     88/100      | 1024/1024  ||    716     |   50.669   |     6      |  3.6e-07   ||\n",
      "||   2023-02-10_06h54m16s    |     89/100      | 1024/1024  ||    716     |   50.669   |     6      |  3.6e-07   ||\n",
      "||   2023-02-10_06h54m17s    |     91/100      | 1024/1024  ||    716     |   50.669   |     6      |  3.6e-07   ||\n",
      "||   2023-02-10_06h54m17s    |     92/100      | 1024/1024  ||    716     |   50.669   |     6      |  3.6e-07   ||\n",
      "||   2023-02-10_06h54m40s    |     93/100      | 1024/1024  ||    716     |   50.669   |     6      |  3.6e-07   ||\n",
      "||   2023-02-10_06h54m41s    |     96/100      | 1024/1024  ||    716     |   50.669   |     6      |  3.6e-07   ||\n",
      "||   2023-02-10_06h55m04s    |     97/100      | 1024/1024  ||    716     |   50.669   |     6      |  3.6e-07   ||\n",
      "||   2023-02-10_06h55m05s    |     98/100      | 1024/1024  ||    716     |   50.669   |     6      |  3.6e-07   ||\n",
      "||   2023-02-10_06h55m05s    |     100/100     | 1024/1024  ||    716     |   50.647   |     7      |  8.5e-07   ||\n",
      "\u001B[0m"
     ]
    }
   ],
   "source": [
    "#============ run the analysis ============\n",
    "!/opt/miniconda3/envs/E2210/bin/python -m nnbits.run --savepath '{savepath}'"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "50f12ae5-4a58-4b2e-bcd3-6ba4dae48bde",
   "metadata": {
    "pycharm": {
     "name": "#%% md\n"
    }
   },
   "source": [
    "# Analyze the results "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "ca98b778-36ee-4974-8a29-96604db34af6",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-06-05T11:46:21.276151Z",
     "start_time": "2022-06-05T11:46:21.273778Z"
    },
    "execution": {
     "iopub.execute_input": "2023-02-10T06:55:40.230408Z",
     "iopub.status.busy": "2023-02-10T06:55:40.229919Z",
     "iopub.status.idle": "2023-02-10T06:55:40.233902Z",
     "shell.execute_reply": "2023-02-10T06:55:40.233454Z",
     "shell.execute_reply.started": "2023-02-10T06:55:40.230378Z"
    },
    "pycharm": {
     "name": "#%%\n"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "\n",
    "# the file-manager is a simple module which knows under which path to search for e.g. the test_accuracies:\n",
    "from nnbits.filemanager import FileManager\n",
    "# given a file-manager, the get_X function loads all the test_accuracies from the analysis above into a single array:\n",
    "from nnbits.bitanalysis import get_X"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "83b9579d-b716-4fd7-baeb-76f70f18c62a",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-06-05T11:46:23.829223Z",
     "start_time": "2022-06-05T11:46:23.825672Z"
    },
    "execution": {
     "iopub.execute_input": "2023-02-10T06:55:40.978424Z",
     "iopub.status.busy": "2023-02-10T06:55:40.978122Z",
     "iopub.status.idle": "2023-02-10T06:55:41.012372Z",
     "shell.execute_reply": "2023-02-10T06:55:41.011808Z",
     "shell.execute_reply.started": "2023-02-10T06:55:40.978395Z"
    },
    "pycharm": {
     "name": "#%%\n"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(100, 1024)"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "F = FileManager('demo_speck32_round7') \n",
    "X = get_X(F)\n",
    "# X contains one row for each of the `N` neural network and one column for each of the `n` bits:\n",
    "X.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "4c4b1c18-4d51-4b21-83fd-f0b6011fc51b",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-06-05T11:46:28.087305Z",
     "start_time": "2022-06-05T11:46:28.084237Z"
    },
    "execution": {
     "iopub.execute_input": "2023-02-10T06:55:43.379196Z",
     "iopub.status.busy": "2023-02-10T06:55:43.378921Z",
     "iopub.status.idle": "2023-02-10T06:55:43.383093Z",
     "shell.execute_reply": "2023-02-10T06:55:43.382644Z",
     "shell.execute_reply.started": "2023-02-10T06:55:43.379177Z"
    },
    "pycharm": {
     "name": "#%%\n"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "# the bit accuracies are calculated by taking the mean over all neural networks which predict the particular bit \n",
    "# (axis=0):\n",
    "# (the following three lines catch a runtime warning if a mean-calculation is empty. This happens if not all bits have been analyzed yet.)\n",
    "import warnings\n",
    "with warnings.catch_warnings():\n",
    "    warnings.simplefilter(\"ignore\", category=RuntimeWarning)\n",
    "    bit_accuracies = np.nanmean(X, axis=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "d283277e-a678-464b-ac92-41384fcee117",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-06-05T11:46:28.612142Z",
     "start_time": "2022-06-05T11:46:28.448131Z"
    },
    "execution": {
     "iopub.execute_input": "2023-02-10T06:55:43.691005Z",
     "iopub.status.busy": "2023-02-10T06:55:43.690752Z",
     "iopub.status.idle": "2023-02-10T06:55:43.833073Z",
     "shell.execute_reply": "2023-02-10T06:55:43.832655Z",
     "shell.execute_reply.started": "2023-02-10T06:55:43.690983Z"
    },
    "pycharm": {
     "name": "#%%\n"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>bit</th>\n",
       "      <th>mean accuracy</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>716</th>\n",
       "      <td>716</td>\n",
       "      <td>0.506469</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>732</th>\n",
       "      <td>732</td>\n",
       "      <td>0.505054</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>525</th>\n",
       "      <td>525</td>\n",
       "      <td>0.501650</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>701</th>\n",
       "      <td>701</td>\n",
       "      <td>0.501575</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>255</th>\n",
       "      <td>255</td>\n",
       "      <td>0.501524</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>19</td>\n",
       "      <td>0.498695</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>302</th>\n",
       "      <td>302</td>\n",
       "      <td>0.498587</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>37</th>\n",
       "      <td>37</td>\n",
       "      <td>0.498459</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>423</th>\n",
       "      <td>423</td>\n",
       "      <td>0.498335</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>290</th>\n",
       "      <td>290</td>\n",
       "      <td>0.498291</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>1024 rows × 2 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "     bit  mean accuracy\n",
       "716  716       0.506469\n",
       "732  732       0.505054\n",
       "525  525       0.501650\n",
       "701  701       0.501575\n",
       "255  255       0.501524\n",
       "..   ...            ...\n",
       "19    19       0.498695\n",
       "302  302       0.498587\n",
       "37    37       0.498459\n",
       "423  423       0.498335\n",
       "290  290       0.498291\n",
       "\n",
       "[1024 rows x 2 columns]"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import pandas as pd\n",
    "df = pd.DataFrame({'bit': np.arange(1024), \n",
    "                   'mean accuracy': bit_accuracies})\n",
    "\n",
    "df.sort_values('mean accuracy', ascending=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "fe4cc4ee-0de5-4eb7-af0a-7e4346aa61d6",
   "metadata": {},
   "source": [
    "# Visualize the Results (shows bit 716 and bit 732)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "c436ec3b-4153-47db-aa1e-71c70cbd465d",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-06-05T11:46:33.835747Z",
     "start_time": "2022-06-05T11:46:33.457403Z"
    },
    "execution": {
     "iopub.execute_input": "2023-02-10T06:55:46.788218Z",
     "iopub.status.busy": "2023-02-10T06:55:46.787943Z",
     "iopub.status.idle": "2023-02-10T06:55:47.754284Z",
     "shell.execute_reply": "2023-02-10T06:55:47.753762Z",
     "shell.execute_reply.started": "2023-02-10T06:55:46.788197Z"
    },
    "pycharm": {
     "name": "#%%\n"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0, 0.5, 'validation accuracy')"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1500x450 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "\n",
    "# here we visualize the obtained `bit_accuracies`:\n",
    "plt.figure(figsize=(10, 3), dpi=150)\n",
    "plt.plot(bit_accuracies, 'o', markersize=3, linestyle = 'None', label=\"mean validation accuracy of each bit\")\n",
    "# find bit with maximum accuracy\n",
    "dfm = df[df['mean accuracy'] == max(df['mean accuracy'])]\n",
    "plt.plot(dfm['bit'], dfm['mean accuracy'], marker='x', c='C0', linestyle = 'None', label=f\"best bit {int(dfm['bit'])} with {float(dfm['mean accuracy'])*100:.1f}% accuracy\")\n",
    "\n",
    "plt.legend(loc='upper left')\n",
    "plt.xlabel('bit id')\n",
    "plt.ylabel('validation accuracy')"
   ]
  }
 ],
 "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": "3.8.12"
  },
  "latex_envs": {
   "LaTeX_envs_menu_present": true,
   "autoclose": false,
   "autocomplete": true,
   "bibliofile": "biblio.bib",
   "cite_by": "apalike",
   "current_citInitial": 1,
   "eqLabelWithNumbers": true,
   "eqNumInitial": 1,
   "hotkeys": {
    "equation": "Ctrl-E",
    "itemize": "Ctrl-I"
   },
   "labels_anchors": false,
   "latex_user_defs": false,
   "report_style_numbering": false,
   "user_envs_cfg": false
  },
  "toc": {
   "base_numbering": 1,
   "nav_menu": {},
   "number_sections": true,
   "sideBar": true,
   "skip_h1_title": false,
   "title_cell": "Table of Contents",
   "title_sidebar": "Contents",
   "toc_cell": true,
   "toc_position": {},
   "toc_section_display": true,
   "toc_window_display": true
  },
  "varInspector": {
   "cols": {
    "lenName": 16,
    "lenType": 16,
    "lenVar": 40
   },
   "kernels_config": {
    "python": {
     "delete_cmd_postfix": "",
     "delete_cmd_prefix": "del ",
     "library": "var_list.py",
     "varRefreshCmd": "print(var_dic_list())"
    },
    "r": {
     "delete_cmd_postfix": ") ",
     "delete_cmd_prefix": "rm(",
     "library": "var_list.r",
     "varRefreshCmd": "cat(var_dic_list()) "
    }
   },
   "types_to_exclude": [
    "module",
    "function",
    "builtin_function_or_method",
    "instance",
    "_Feature"
   ],
   "window_display": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}