lr_scheduler.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "\n",
    "import math"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [],
   "source": [
    "learning_rate = 5e-4\n",
    "warmup_iters = 0\n",
    "\n",
    "lr_decay_iters = 10000 #max_iters\n",
    "min_lr = 5e-5"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [],
   "source": [
    "# learning rate decay scheduler (cosine with warmup)\n",
    "def get_lr(it):\n",
    "    # 1) linear warmup for warmup_iters steps\n",
    "    if lr_decay_iters == 0:\n",
    "        return learning_rate\n",
    "    \n",
    "    if it < warmup_iters:\n",
    "        return learning_rate * it / warmup_iters\n",
    "    # 2) if it > lr_decay_iters, return min learning rate\n",
    "    if it > lr_decay_iters:\n",
    "        return min_lr\n",
    "    # 3) in between, use cosine decay down to min learning rate\n",
    "    decay_ratio = (it - warmup_iters) / (lr_decay_iters - warmup_iters)\n",
    "    assert 0 <= decay_ratio <= 1\n",
    "    coeff = 0.5 * (1.0 + math.cos(math.pi * decay_ratio))  # coeff ranges 0..1\n",
    "    return min_lr + coeff * (learning_rate - min_lr)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [],
   "source": [
    "lrs = [get_lr(it) for it in range(20000)]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x7f9191512ed0>]"
      ]
     },
     "execution_count": 48,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(lrs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "torch23",
   "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": "3.11.3"
  },
  "orig_nbformat": 4
 },
 "nbformat": 4,
 "nbformat_minor": 2
}