{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of samples:  [ 16.  32.  64. 128. 256.] (x2)\n",
      "Number of samples:  [  8.  16.  32.  64. 128.] (x4)\n",
      "Number of samples:  [ 4.  8. 16. 32. 64.] (x14)\n",
      "Number of samples:  [ 4.  8. 16. 32. 64.] (x20)\n",
      "Number of samples:  [ 2.  4.  8. 16. 32.] (x77)\n"
     ]
    }
   ],
   "source": [
    "# This is equivalent to round factor / log2(num_classes) to the nearest power of 2\n",
    "factor = np.array([16, 32, 64, 128, 256])\n",
    "for num_classes in [2, 4, 14, 20, 77]:\n",
    "    scale = factor / np.log2(num_classes)\n",
    "    nearest_power_of_2 = 2 ** np.round(np.log2(scale)) # round to nearest power of 2\n",
    "    print(\"Number of samples: \", nearest_power_of_2, f\"(x{num_classes})\")"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "llmcal",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.11.10"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}