mix_2_domains.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch\n",
    "from collections import defaultdict\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "from law import ScalingLaw\n",
    "\n",
    "def load_loss(path):\n",
    "    losses = defaultdict(lambda: defaultdict(dict))   # domain -> ratio\n",
    "    with open(path, \"r\") as f:\n",
    "        lines = set(f.readlines())\n",
    "    for line in lines:   \n",
    "        model, data, loss = line.split()\n",
    "        model = model.strip('/')\n",
    "        step = int(model.split('/')[-1])\n",
    "        ratio = \"-\".join(model.split('/')[-2].split('-')[1:])\n",
    "        subdomain = data.split('/')[-1].split('.')[-2]\n",
    "        losses[ratio][subdomain][step] = float(loss)\n",
    "    return losses\n",
    "\n",
    "LOSSES = {\n",
    "    \"70M\": load_loss(\"data/70M.txt\"),\n",
    "    \"160M\": load_loss(\"data/160M.txt\"),\n",
    "}   # sz, ratio, domain, step"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "ys = dict()\n",
    "for i in [\"70M\", \"160M\"]:\n",
    "    losses = LOSSES[i]\n",
    "    ratios = list(losses.keys())\n",
    "    domains = list(losses[ratios[0]].keys())\n",
    "    steps = list(losses[ratios[0]][domains[0]].keys())\n",
    "    ratios = np.sort(ratios)\n",
    "    ys[i] = defaultdict(lambda:defaultdict(list))\n",
    "    for r in ratios:\n",
    "        for d in [\"Github\", \"Pile-CC\"]:\n",
    "            for s in losses[r][d].keys():\n",
    "                ys[i][d][s].append(losses[r][d][s])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "def inter_law(x, param):\n",
    "    k, b = param\n",
    "    # y = c + exp(kx+b)\n",
    "    return torch.exp(k*x + b)\n",
    "\n",
    "def param_generator():\n",
    "    for k in np.linspace(-2.4, -1.6, 11):\n",
    "        for b in np.linspace(-1.0, -0.1, 11):\n",
    "            yield [k, b]\n",
    "###\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "x0 = [float(r.split('-')[0]) for r in ratios]\n",
    "x1 = [float(r.split('-')[1]) for r in ratios]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/tmp/ipykernel_130051/759608647.py:9: RuntimeWarning: divide by zero encountered in log\n",
      "  corr = np.abs(np.corrcoef(x, np.log(y-c))[0, 1])\n",
      "/cpfs01/projects-HDD/cfff-173661e84712_HDD/public/miniconda/envs/sl/lib/python3.11/site-packages/numpy/lib/function_base.py:2742: RuntimeWarning: invalid value encountered in subtract\n",
      "  X -= avg[:, None]\n",
      "100%|█████████████████████████████████████████████████████████████████████████████████████████████| 121/121 [00:04<00:00, 25.85it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1.0180197932641022e-05 99\n",
      "tensor([-2.4381, -0.5978])\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|█████████████████████████████████████████████████████████████████████████████████████████████| 121/121 [00:04<00:00, 28.92it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "6.930810741323512e-06 99\n",
      "tensor([-1.5473, -0.6506])\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|█████████████████████████████████████████████████████████████████████████████████████████████| 121/121 [00:04<00:00, 27.34it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "3.6705273487314116e-06 99\n",
      "tensor([-1.6405, -0.6819])\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|█████████████████████████████████████████████████████████████████████████████████████████████| 121/121 [00:03<00:00, 33.36it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "9.003677519103803e-07 99\n",
      "tensor([-2.9041, -0.5333])\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|█████████████████████████████████████████████████████████████████████████████████████████████| 121/121 [00:04<00:00, 27.10it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2.7243918339081574e-06 99\n",
      "tensor([-2.2959, -0.5862])\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|█████████████████████████████████████████████████████████████████████████████████████████████| 121/121 [00:03<00:00, 31.54it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1.6656942989357049e-06 99\n",
      "tensor([-2.5189, -0.6257])\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|█████████████████████████████████████████████████████████████████████████████████████████████| 121/121 [00:04<00:00, 27.05it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "8.977600373327732e-05 99\n",
      "tensor([-3.6654, -0.1843])\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|█████████████████████████████████████████████████████████████████████████████████████████████| 121/121 [00:04<00:00, 28.92it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "4.9896952987182885e-06 99\n",
      "tensor([-1.6850, -0.7528])\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|█████████████████████████████████████████████████████████████████████████████████████████████| 121/121 [00:03<00:00, 31.01it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2.7009764380636625e-06 99\n",
      "tensor([-1.6139, -0.7760])\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|█████████████████████████████████████████████████████████████████████████████████████████████| 121/121 [00:03<00:00, 31.06it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.0003784890868701041 99\n",
      "tensor([-0.1236,  1.3070])\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|█████████████████████████████████████████████████████████████████████████████████████████████| 121/121 [00:04<00:00, 27.06it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "3.5466832741803955e-06 99\n",
      "tensor([-2.4178, -0.5877])\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|█████████████████████████████████████████████████████████████████████████████████████████████| 121/121 [00:04<00:00, 24.46it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "3.302324785181554e-06 99\n",
      "tensor([-2.7239, -0.6131])\n"
     ]
    }
   ],
   "source": [
    "params = defaultdict(lambda:defaultdict(lambda:defaultdict(list)))\n",
    "for m in [\"70M\", \"160M\"]:\n",
    "    for d in domains:\n",
    "        for s in [10000, 20000, 30000]:\n",
    "            y = ys[m][d][s]\n",
    "            x = x0 if d == \"Github\" else x1\n",
    "            best_c, max_corr = 0, 0\n",
    "            for c in np.linspace(0, min(y), 1000):\n",
    "                corr = np.abs(np.corrcoef(x, np.log(y-c))[0, 1])\n",
    "                if corr > max_corr:\n",
    "                    max_corr = corr\n",
    "                    best_c = c\n",
    "            law = ScalingLaw(inter_law)\n",
    "            p = law.fit(x, y-best_c, param_generator(), max_step=100, delta=0.02)\n",
    "            params[m][d][s] = [best_c, p[0], p[1]] # param\n",
    "        "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 800x400 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from matplotlib.cm import get_cmap\n",
    "import matplotlib.patches as mpatches\n",
    "import matplotlib\n",
    "\n",
    "plt.rcParams['font.family'] = 'Helvetica'  # Set default font family\n",
    "\n",
    "\n",
    "c1 = np.array([[253, 191, 96], [255, 137, 17]])/255\n",
    "c2 = np.array([[159, 112, 253], [127, 39, 255]])/255\n",
    "c = np.array([c1, c2])\n",
    "\n",
    "fig, axes = plt.subplots(1, 2, figsize=(8, 4))\n",
    "fig.subplots_adjust(bottom=0.2, top=1, wspace)\n",
    "\n",
    "xs = np.linspace(0, 1, 20)\n",
    "\n",
    "xx = [x0, x1]\n",
    "for k, d in enumerate([\"Github\", \"Pile-CC\"]):\n",
    "    x = xx[k]\n",
    "    ax = axes[k]\n",
    "\n",
    "    for i, m in enumerate([\"70M\", \"160M\"]):\n",
    "        for j, s in enumerate([20000, 30000]):\n",
    "            y = np.array(ys[m][d][s])\n",
    "            p = params[m][d][s]\n",
    "            ax.scatter(x, y-p[0], color=c[i][j], s=80)\n",
    "            ax.plot(xs, inter_law(torch.tensor(xs), torch.tensor(p[1:])), color=c[i][j])\n",
    "            ax.set_yscale(\"log\")\n",
    "            # ax.legend()\n",
    "            ax.set_xlabel(f\"Proportion of {d}\", fontsize=14)\n",
    "            ax.set_ylabel(f\"Reducible Loss on {d}\", fontsize=14)\n",
    "            ax.spines['right'].set_visible(False)\n",
    "            ax.spines['top'].set_visible(False)\n",
    "legend_elements = []\n",
    "for i, m in enumerate([\"70M\", \"160M\"]):\n",
    "    for j, s in enumerate([20000, 30000]):\n",
    "        el = mpatches.Patch(color=c[i][j], label=f\"{m} / {s//1000}k steps\")\n",
    "        legend_elements.append(el)\n",
    "    \n",
    "# Create the legend and place it at the bottom of the figure\n",
    "fig.legend(handles=legend_elements, loc='lower center', bbox_to_anchor=(0.47, 0), ncol=4)\n",
    "# plt.tight_layout()\n",
    "plt.savefig('savefigs/2mix.png', dpi=300)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "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": "3.11.0"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}