Add files using upload-large-folder tool

.gitattributes

@@ -56,3 +56,6 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 # Video files - compressed
 *.mp4 filter=lfs diff=lfs merge=lfs -text
 *.webm filter=lfs diff=lfs merge=lfs -text
+SPICE_2_dipeptides_all_configs_test.xyz filter=lfs diff=lfs merge=lfs -text
+SPICE_2_dipeptides_all_configs_train.xyz filter=lfs diff=lfs merge=lfs -text
+maceoff_1_with_multipoles_test.xyz filter=lfs diff=lfs merge=lfs -text

SPICE_2_dipeptides_all_configs_test.xyz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:40a5031f19ab578ec3f3fe25fb01fd0c9f2ee3f603405b522d8598e641faee80
+size 58054596

SPICE_2_dipeptides_all_configs_train.xyz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f9fc680822419a4e56d30accf646c84d654ef9cb4bdab2d7086f9f29af84925c
+size 233004549

analysis.ipynb

@@ -0,0 +1,189 @@
+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from ase.io import read, write"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "train_configs = read(\"maceoff_1_with_multipoles_train.xyz\", \":\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 3,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "dict_keys(['numbers', 'positions', 'formal_charges', 'mbis_multipoles', 'forces'])"
+      ]
+     },
+     "execution_count": 3,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "train_configs[0].arrays.keys()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 5,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "image/png": "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",
+      "text/plain": [
+       "<Figure size 640x480 with 1 Axes>"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    }
+   ],
+   "source": [
+    "import numpy as np\n",
+    "import matplotlib.pyplot as plt\n",
+    "max_forces = [np.linalg.norm(c.arrays[\"forces\"], axis=1).max(axis=0) for c in train_configs]\n",
+    "\n",
+    "\n",
+    "plt.yscale(\"log\")\n",
+    "plt.hist(max_forces, bins=100)\n",
+    "plt.show()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 11,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "2246"
+      ]
+     },
+     "execution_count": 11,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "bad_configs = [c for (c, f) in zip(train_configs, max_forces) if f > 15]\n",
+    "len(bad_configs)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 12,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "write(\"bad_configs.xyz\", bad_configs)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 13,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "943200"
+      ]
+     },
+     "execution_count": 13,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "good_configs = [c for (c, f) in zip(train_configs, max_forces) if f <= 15]\n",
+    "len(good_configs)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 14,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "/opt/conda/envs/mace-openmm/lib/python3.11/site-packages/ase/io/extxyz.py:1000: UserWarning: write_xyz() overwriting array \"forces\" present in atoms.arrays with stored results from calculator\n",
+      "  warnings.warn('write_xyz() overwriting array \"{0}\" present '\n"
+     ]
+    }
+   ],
+   "source": [
+    "write(\"maceoff_1_with_multipoles_train_force_filter_15.xyz\", good_configs)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 15,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "image/png": "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",
+      "text/plain": [
+       "<Figure size 640x480 with 1 Axes>"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    }
+   ],
+   "source": [
+    "charges = [np.sum(c.arrays[\"mbis_multipoles\"][:,0]) for c in good_configs]\n",
+    "\n",
+    "plt.hist(charges, bins=100)\n",
+    "plt.show()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "mace-openmm",
+   "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": "3.11.9"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 2
+}

analysis_spice2.ipynb

@@ -0,0 +1,199 @@
+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from ase.io import read, write"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "train_configs = read(\"SPICE_v2_train.xyz\", \":\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 3,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "dict_keys(['numbers', 'positions', 'formal_charges', 'mbis_multipoles', 'forces'])"
+      ]
+     },
+     "execution_count": 3,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "train_configs[0].arrays.keys()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 4,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "image/png": "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",
+      "text/plain": [
+       "<Figure size 640x480 with 1 Axes>"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    }
+   ],
+   "source": [
+    "import numpy as np\n",
+    "import matplotlib.pyplot as plt\n",
+    "max_forces = [np.linalg.norm(c.arrays[\"forces\"], axis=1).max(axis=0) for c in train_configs]\n",
+    "\n",
+    "\n",
+    "plt.yscale(\"log\")\n",
+    "plt.hist(max_forces, bins=100)\n",
+    "plt.show()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 5,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "28"
+      ]
+     },
+     "execution_count": 5,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "bad_configs = [c for (c, f) in zip(train_configs, max_forces) if f > 15]\n",
+    "len(bad_configs)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 6,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "/opt/conda/envs/mace-scf/lib/python3.11/site-packages/ase/io/extxyz.py:1000: UserWarning: write_xyz() overwriting array \"forces\" present in atoms.arrays with stored results from calculator\n",
+      "  warnings.warn('write_xyz() overwriting array \"{0}\" present '\n"
+     ]
+    }
+   ],
+   "source": [
+    "write(\"bad_configs.xyz\", bad_configs)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 7,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "1428672"
+      ]
+     },
+     "execution_count": 7,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "good_configs = [c for (c, f) in zip(train_configs, max_forces) if f <= 15]\n",
+    "len(good_configs)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 14,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "/opt/conda/envs/mace-openmm/lib/python3.11/site-packages/ase/io/extxyz.py:1000: UserWarning: write_xyz() overwriting array \"forces\" present in atoms.arrays with stored results from calculator\n",
+      "  warnings.warn('write_xyz() overwriting array \"{0}\" present '\n"
+     ]
+    }
+   ],
+   "source": [
+    "write(\"spice_v2_15_eV-A_force_filter.xyz\", good_configs)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 9,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAicAAAGdCAYAAADJ6dNTAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjkuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8hTgPZAAAACXBIWXMAAA9hAAAPYQGoP6dpAAAf5ElEQVR4nO3df2xddf3H8ddtZ283aa8rhduVdSsQQQt4q/3FdPPbkhtqgcZNxSUauGtII8kt0dyg6cS0/jEoETIb8GgRMyZGQoPGTR1pMsqPTizp1lIUliGNnanM3nYOeteLacft/f5hvFi7jd71rudzep6P5P5xzz07592jsU/v/ZxbTzKZTAoAAMAQWXYPAAAA8N+IEwAAYBTiBAAAGIU4AQAARiFOAACAUYgTAABgFOIEAAAYhTgBAABGWWX3AOmam5vTiRMnlJeXJ4/HY/c4AABgEZLJpE6fPq3i4mJlZZ3/vRHHxcmJEydUUlJi9xgAAOACjI2Naf369efdx3FxkpeXJ+nfP1x+fr7N0wAAgMWIxWIqKSlJ/R4/H8fFyX8+ysnPzydOAABwmMUsyWBBLAAAMApxAgAAjEKcAAAAozgmTizLUllZmaqqquweBQAAXESeZDKZtHuIdMRiMfl8Pk1NTbEgFgAAh0jn97dj3jkBAADuQJwAAACjECcAAMAoxAkAADAKcQIAAIxCnAAAAKMQJwAAwCjECQAAMIrj/ioxAGB5lLYeWLDt+IO32jAJ3IZ3TgAAgFFsiZPR0VHV1dWprKxMN9xwg+LxuB1jAAAAA9nysc6OHTu0a9cubdmyRadOnZLX67VjDAAAYKBlj5M33nhDH/nIR7RlyxZJUkFBwXKPAAAADJb2xzp9fX1qbGxUcXGxPB6P9u3bt2Afy7JUWlqq3Nxc1dTUaGBgIPXaW2+9pUsuuUSNjY36zGc+owceeGBJPwAAAFhZ0o6TeDyuQCAgy7LO+np3d7cikYja29s1NDSkQCCg+vp6TUxMSJLef/99HTp0SD/+8Y/V39+vgwcP6uDBg+c838zMjGKx2LwHAABYudKOk4aGBu3atUvbtm076+u7d+9Wc3OzmpqaVFZWpq6uLq1Zs0Z79uyRJF1xxRWqrKxUSUmJvF6vbrnlFg0PD5/zfB0dHfL5fKlHSUlJuiMDAAAHyejdOrOzsxocHFQwGPzgBFlZCgaD6u/vlyRVVVVpYmJC77zzjubm5tTX16dPfvKT5zzmzp07NTU1lXqMjY1lcmQAAGCYjC6IPXnypBKJhPx+/7ztfr9fx44d+/cJV63SAw88oM9//vNKJpO6+eabddttt53zmF6vl7t5AABwEVtuJW5oaFBDQ0Na/8ayLFmWpUQicZGmAgAAJsjoxzqFhYXKzs5WNBqdtz0ajaqoqGhJxw6Hwzp69KgOHz68pOMAAACzZTROcnJyVFFRod7e3tS2ubk59fb2atOmTZk8FQAAWKHS/lhnenpaIyMjqeejo6MaHh5WQUGBNmzYoEgkolAopMrKSlVXV6uzs1PxeFxNTU1LGpSPdQAAcAdPMplMpvMPXnzxRdXV1S3YHgqFtHfvXknSj370Iz300EMaHx9XeXm5HnnkEdXU1GRk4FgsJp/Pp6mpKeXn52fkmACAhfirxMikdH5/px0ndiNOAGB5ECfIpHR+f9vyV4kBAADOxTFxYlmWysrKVFVVZfcoAADgInJMnHArMQAA7uCYOAEAAO5AnAAAAKM4Jk5YcwIAgDs4Jk5YcwIAgDs4Jk4AAIA7ECcAAMAojokT1pwAAOAOjokT1pwAAOAOjokTAADgDsQJAAAwCnECAACMQpwAAACjOCZOuFsHAAB3cEyccLcOAADu4Jg4AQAA7kCcAAAAoxAnAADAKMQJAAAwCnECAACMQpwAAACjOCZO+J4TAADcwTFxwvecAADgDo6JEwAA4A7ECQAAMApxAgAAjEKcAAAAoxAnAADAKMQJAAAwCnECAACMQpwAAACjOCZO+IZYAADcwTFxwjfEAgDgDo6JEwAA4A7ECQAAMApxAgAAjEKcAAAAoxAnAADAKMQJAAAwCnECAACMQpwAAACjECcAAMAoxAkAADDKKjtOWlpaqvz8fGVlZWnt2rV64YUX7BgDAAAYyJY4kaQ//vGPuuSSS+w6PQAAMBQf6wAAAKOkHSd9fX1qbGxUcXGxPB6P9u3bt2Afy7JUWlqq3Nxc1dTUaGBgYN7rHo9H//d//6eqqir98pe/vODhAQDAypN2nMTjcQUCAVmWddbXu7u7FYlE1N7erqGhIQUCAdXX12tiYiK1zx/+8AcNDg7qt7/9rR544AH96U9/uvCfAAAArChpx0lDQ4N27dqlbdu2nfX13bt3q7m5WU1NTSorK1NXV5fWrFmjPXv2pPa54oorJEnr1q3TLbfcoqGhoXOeb2ZmRrFYbN4DAACsXBldczI7O6vBwUEFg8EPTpCVpWAwqP7+fkn/fufl9OnTkqTp6Wk9//zzuu666855zI6ODvl8vtSjpKQkkyMDAADDZDROTp48qUQiIb/fP2+73+/X+Pi4JCkajWrz5s0KBAK68cYbdeedd6qqquqcx9y5c6empqZSj7GxsUyODAAADLPstxJfddVVeu211xa9v9frldfrvYgTAQAAk2T0nZPCwkJlZ2crGo3O2x6NRlVUVLSkY1uWpbKysvO+ywIAAJwvo3GSk5OjiooK9fb2prbNzc2pt7dXmzZtWtKxw+Gwjh49qsOHDy91TAAAYLC0P9aZnp7WyMhI6vno6KiGh4dVUFCgDRs2KBKJKBQKqbKyUtXV1ers7FQ8HldTU1NGBwcAACtT2nFy5MgR1dXVpZ5HIhFJUigU0t69e7V9+3ZNTk6qra1N4+PjKi8vV09Pz4JFsumyLEuWZSmRSCzpOAAAwGyeZDKZtHuIdMRiMfl8Pk1NTSk/P9/ucQBgxSptPbBg2/EHb7VhEqwE6fz+5m/rAAAAozgmTrhbBwAAd3BMnHC3DgAA7uCYOAEAAO6w7N8QCwBApvzvol0W7K4MjnnnhDUnAAC4g2PihDUnAAC4g2PiBAAAuANxAgAAjEKcAAAAozjmbh3+tg6AlYY7TYCzc8w7JyyIBQDAHRwTJwAAwB2IEwAAYBTiBAAAGIU4AQAARnFMnPD19QAAuINj4oS7dQAAcAfHxAkAAHAH4gQAABiFOAEAAEYhTgAAgFGIEwAAYBTHxAm3EgMA4A6OiRNuJQYAwB0cEycAAMAdiBMAAGAU4gQAABiFOAEAAEYhTgAAgFGIEwAAYBTiBAAAGIU4AQAARiFOAACAURwTJ3x9PQAA7uCYOOHr6wEAcAfHxAkAAHAH4gQAABiFOAEAAEYhTgAAgFGIEwAAYBTiBAAAGIU4AQAARiFOAACAUYgTAABgFOIEAAAYxbY4ee+997Rx40bde++9do0AAAAMZFuc3H///brxxhvtOj0AADCULXHy1ltv6dixY2poaLDj9AAAwGBpx0lfX58aGxtVXFwsj8ejffv2LdjHsiyVlpYqNzdXNTU1GhgYmPf6vffeq46OjgseGgAArFxpx0k8HlcgEJBlWWd9vbu7W5FIRO3t7RoaGlIgEFB9fb0mJiYkSfv379c111yja665ZmmTAwCAFWlVuv+goaHhvB/H7N69W83NzWpqapIkdXV16cCBA9qzZ49aW1v1yiuv6Omnn9Yzzzyj6elpnTlzRvn5+Wprazvr8WZmZjQzM5N6HovF0h0ZAAA4SEbXnMzOzmpwcFDBYPCDE2RlKRgMqr+/X5LU0dGhsbExHT9+XA8//LCam5vPGSb/2d/n86UeJSUlmRwZAAAYJqNxcvLkSSUSCfn9/nnb/X6/xsfHL+iYO3fu1NTUVOoxNjaWiVEBAICh0v5YJ5N27Njxoft4vV55vV5ZliXLspRIJC7+YAAAwDYZfeeksLBQ2dnZikaj87ZHo1EVFRUt6djhcFhHjx7V4cOHl3QcAABgtozGSU5OjioqKtTb25vaNjc3p97eXm3atCmTpwIAACtU2h/rTE9Pa2RkJPV8dHRUw8PDKigo0IYNGxSJRBQKhVRZWanq6mp1dnYqHo+n7t65UHysAwCAO6QdJ0eOHFFdXV3qeSQSkSSFQiHt3btX27dv1+TkpNra2jQ+Pq7y8nL19PQsWCSbrnA4rHA4rFgsJp/Pt6RjAQAAc6UdJ7W1tUomk+fdp6WlRS0tLRc8FAAAcC/b/vAfAADA2TgmTizLUllZmaqqquweBQAAXESOiRNuJQYAwB0cEycAAMAdiBMAAGAUx8QJa04AAHAHx8QJa04AAHAHx8QJAABwB+IEAAAYhTgBAABGcUycsCAWAAB3cEycsCAWAAB3cEycAAAAdyBOAACAUYgTAABgFMfECQtiAQBwB8fECQtiAQBwB8fECQAAcAfiBAAAGIU4AQAARiFOAACAUYgTAABgFOIEAAAYxTFxwvecAADgDqvsHmCxwuGwwuGwYrGYfD6f3eMAMFxp64EF244/eKsNkwBIl2PeOQEAAO5AnAAAAKMQJwAAwCjECQAAMApxAgAAjOKYu3UAAHAa7hq7MLxzAgAAjEKcAAAAozgmTviGWAAA3MExcRIOh3X06FEdPnzY7lEAAMBF5Jg4AQAA7kCcAAAAoxAnAADAKMQJAAAwCnECAACMQpwAAACjECcAAMAoxAkAADAKcQIAAIxCnAAAAKMse5y8++67qqysVHl5ua6//no9/vjjyz0CAAAw2KrlPmFeXp76+vq0Zs0axeNxXX/99frSl76kSy+9dLlHAQAABlr2d06ys7O1Zs0aSdLMzIySyaSSyeRyjwEAAAyVdpz09fWpsbFRxcXF8ng82rdv34J9LMtSaWmpcnNzVVNTo4GBgXmvv/vuuwoEAlq/fr2+/e1vq7Cw8IJ/AAAAsLKkHSfxeFyBQECWZZ319e7ubkUiEbW3t2toaEiBQED19fWamJhI7fOxj31Mr732mkZHR/XUU08pGo1e+E8AAABWlLTjpKGhQbt27dK2bdvO+vru3bvV3NyspqYmlZWVqaurS2vWrNGePXsW7Ov3+xUIBHTo0KFznm9mZkaxWGzeAwAArFwZXXMyOzurwcFBBYPBD06QlaVgMKj+/n5JUjQa1enTpyVJU1NT6uvr07XXXnvOY3Z0dMjn86UeJSUlmRwZAAAYJqNxcvLkSSUSCfn9/nnb/X6/xsfHJUl/+9vftGXLFgUCAW3ZskX33HOPbrjhhnMec+fOnZqamko9xsbGMjkyAAAwzLLfSlxdXa3h4eFF7+/1euX1ei/eQAAAwCgZfeeksLBQ2dnZCxa4RqNRFRUVLenYlmWprKxMVVVVSzoOAAAwW0bjJCcnRxUVFert7U1tm5ubU29vrzZt2rSkY4fDYR09elSHDx9e6pgAAMBgaX+sMz09rZGRkdTz0dFRDQ8Pq6CgQBs2bFAkElEoFFJlZaWqq6vV2dmpeDyupqamjA4OAABWprTj5MiRI6qrq0s9j0QikqRQKKS9e/dq+/btmpycVFtbm8bHx1VeXq6enp4Fi2TTZVmWLMtSIpFY0nEAAIDZ0o6T2traD/26+ZaWFrW0tFzwUGcTDocVDocVi8Xk8/kyemwAAGCOZf/bOgAAAOez7LcSXyg+1gHMUtp6YN7z4w/eatMkAFYax7xzwt06AAC4g2PiBAAAuANxAgAAjOKYOOEbYgEAcAfHxAlrTgAAcAfHxAkAAHAH4gQAABiFOAEAAEZxTJywIBYAAHdwTJywIBYAAHdwTJwAAAB3IE4AAIBRiBMAAGAU4gQAABjFMXHC3ToAALiDY+KEu3UAAHAHx8QJAABwh1V2DwBgodLWA/OeH3/wVpsmAYDlxzsnAADAKMQJAAAwCnECAACM4pg44VZiAADcwTFxwq3EAAC4g2PiBAAAuANxAgAAjEKcAAAAoxAnAADAKMQJAAAwCnECAACMQpwAAACjECcAAMAoxAkAADCKY+KEr68HAMAdHBMnfH09AADu4Jg4AQAA7kCcAAAAoxAnAADAKMQJAAAwCnECAACMQpwAAACjECcAAMAoxAkAADAKcQIAAIxCnAAAAKMse5yMjY2ptrZWZWVl+tSnPqVnnnlmuUcAAAAGW7XsJ1y1Sp2dnSovL9f4+LgqKip0yy236KMf/ehyjwIAAAy07HGybt06rVu3TpJUVFSkwsJCnTp1ijgBAACSLuBjnb6+PjU2Nqq4uFgej0f79u1bsI9lWSotLVVubq5qamo0MDBw1mMNDg4qkUiopKQk7cEBAMDKlHacxONxBQIBWZZ11te7u7sViUTU3t6uoaEhBQIB1dfXa2JiYt5+p06d0p133qmf/vSnFzY5AABYkdL+WKehoUENDQ3nfH337t1qbm5WU1OTJKmrq0sHDhzQnj171NraKkmamZnR1q1b1draqs9+9rPnPd/MzIxmZmZSz2OxWLojAwAAB8no3Tqzs7MaHBxUMBj84ARZWQoGg+rv75ckJZNJ7dixQzfddJPuuOOODz1mR0eHfD5f6sFHQAAArGwZjZOTJ08qkUjI7/fP2+73+zU+Pi5Jevnll9Xd3a19+/apvLxc5eXl+vOf/3zOY+7cuVNTU1Opx9jYWCZHBgAAhln2u3U2b96subm5Re/v9Xrl9XplWZYsy1IikbiI0wEAALtl9J2TwsJCZWdnKxqNztsejUZVVFS0pGOHw2EdPXpUhw8fXtJxAACA2TIaJzk5OaqoqFBvb29q29zcnHp7e7Vp06ZMngoAAKxQaX+sMz09rZGRkdTz0dFRDQ8Pq6CgQBs2bFAkElEoFFJlZaWqq6vV2dmpeDyeunsHK19p64F5z48/eKtNk5zd/84nmTcjALhZ2nFy5MgR1dXVpZ5HIhFJUigU0t69e7V9+3ZNTk6qra1N4+PjKi8vV09Pz4JFsulizQkAAO6QdpzU1tYqmUyed5+Wlha1tLRc8FBnEw6HFQ6HFYvF5PP5MnpsAABgjmX/q8QAAADn45g4sSxLZWVlqqqqsnsUAABwETkmTriVGAAAd3BMnAAAAHcgTgAAgFEcEyesOQEAwB0cEyesOQEAwB0cEycAAMAdiBMAAGAU4gQAABjFMXHCglgAANzBMXHCglgAANzBMXECAADcgTgBAABGIU4AAIBRHBMnLIgFAMAdHBMnLIgFAMAdHBMnAADAHVbZPYBpSlsPzHt+/MFbbZoEAAB34p0TAABgFOIEAAAYhTgBAABGIU4AAIBRHBMnfM8JAADu4Jg44XtOAABwB8fECQAAcAfiBAAAGIU4AQAARiFOAACAUYgTAABgFOIEAAAYhTgBAABGIU4AAIBRHBMnfEMsAADu4Jg44RtiAQBwB8fECQAAcAfiBAAAGIU4AQAARiFOAACAUYgTAABgFOIEAAAYhTgBAABGIU4AAIBRiBMAAGAU4gQAABjFljjZtm2b1q5dq6985St2nB4AABjMljj55je/qSeffNKOUwMAAMPZEie1tbXKy8uz49QAAMBwacdJX1+fGhsbVVxcLI/Ho3379i3Yx7IslZaWKjc3VzU1NRoYGMjErAAAwAXSjpN4PK5AICDLss76end3tyKRiNrb2zU0NKRAIKD6+npNTEwseVgAALDyrUr3HzQ0NKihoeGcr+/evVvNzc1qamqSJHV1denAgQPas2ePWltb0x5wZmZGMzMzqeexWCztYwAAAOfI6JqT2dlZDQ4OKhgMfnCCrCwFg0H19/df0DE7Ojrk8/lSj5KSkkyNCwAADJTRODl58qQSiYT8fv+87X6/X+Pj46nnwWBQt99+u5599lmtX7/+vOGyc+dOTU1NpR5jY2OZHBkAABgm7Y91MuG5555b9L5er1der/ciTgMAAEyS0XdOCgsLlZ2drWg0Om97NBpVUVHRko5tWZbKyspUVVW1pOMAAACzZTROcnJyVFFRod7e3tS2ubk59fb2atOmTUs6djgc1tGjR3X48OGljgkAAAyW9sc609PTGhkZST0fHR3V8PCwCgoKtGHDBkUiEYVCIVVWVqq6ulqdnZ2Kx+Opu3cAAADOJ+04OXLkiOrq6lLPI5GIJCkUCmnv3r3avn27Jicn1dbWpvHxcZWXl6unp2fBItl0WZYly7KUSCSWdBwAAGC2tOOktrZWyWTyvPu0tLSopaXlgoc6m3A4rHA4rFgsJp/Pl9FjAwAAc9jyt3UAAADOhTgBAABGcUyccCsxAADu4Jg44VZiAADcwTFxAgAA3IE4AQAARnFMnLDmBAAAd3BMnLDmBAAAd3BMnAAAAHcgTgAAgFEcEyesOQEAwB0cEyesOQEAwB0cEycAAMAdiBMAAGAU4gQAABiFOAEAAEZxTJxwtw4AAO7gmDjhbh0AANzBMXECAADcgTgBAABGIU4AAIBRiBMAAGAU4gQAABhlld0DLJZlWbIsS4lEwu5RAABYMUpbDyzYdvzBW22Y5AOOeeeEW4kBAHAHx8QJAABwB+IEAAAYhTgBAABGIU4AAIBRiBMAAGAU4gQAABiFOAEAAEYhTgAAgFGIEwAAYBTHxIllWSorK1NVVZXdowAAgIvIMXHC19cDAOAOjokTAADgDsQJAAAwCnECAACMssruAdKVTCYlSbFY7KIcf27mvXnPL9Z5VjLTr+H/zieZP6Np80nmz8h/zkvHNVw6J15D6eLM+J9j/uf3+Pl4kovZyyB///vfVVJSYvcYAADgAoyNjWn9+vXn3cdxcTI3N6cTJ04oLy9PHo8no8eOxWIqKSnR2NiY8vPzM3psN+E6ZgbXMTO4jpnBdcwMN1/HZDKp06dPq7i4WFlZ519V4riPdbKysj60uJYqPz/fdf+luRi4jpnBdcwMrmNmcB0zw63X0efzLWo/FsQCAACjECcAAMAoxMl/8Xq9am9vl9frtXsUR+M6ZgbXMTO4jpnBdcwMruPiOG5BLAAAWNl45wQAABiFOAEAAEYhTgAAgFGIEwAAYBTi5Bz+8pe/6Itf/KIKCwuVn5+vzZs364UXXrB7LEc6cOCAampqtHr1aq1du1Zbt261eyTHmpmZUXl5uTwej4aHh+0ex1GOHz+uu+66S1deeaVWr16tq6++Wu3t7ZqdnbV7NONZlqXS0lLl5uaqpqZGAwMDdo/kKB0dHaqqqlJeXp4uv/xybd26VW+++abdYxmNODmH2267Te+//76ef/55DQ4OKhAI6LbbbtP4+LjdoznKr3/9a91xxx1qamrSa6+9ppdffllf+9rX7B7Lsb7zne+ouLjY7jEc6dixY5qbm9Njjz2mN954Qz/84Q/V1dWl7373u3aPZrTu7m5FIhG1t7draGhIgUBA9fX1mpiYsHs0x3jppZcUDof1yiuv6ODBgzpz5oxuvvlmxeNxu0czVxILTE5OJiUl+/r6UttisVhSUvLgwYM2TuYsZ86cSV5xxRXJn/3sZ3aPsiI8++yzyU984hPJN954Iykp+eqrr9o9kuP94Ac/SF555ZV2j2G06urqZDgcTj1PJBLJ4uLiZEdHh41TOdvExERSUvKll16yexRj8c7JWVx66aW69tpr9eSTTyoej+v999/XY489pssvv1wVFRV2j+cYQ0NDevvtt5WVlaVPf/rTWrdunRoaGvT666/bPZrjRKNRNTc36xe/+IXWrFlj9zgrxtTUlAoKCuwew1izs7MaHBxUMBhMbcvKylIwGFR/f7+Nkznb1NSUJPHfvfMgTs7C4/Houeee06uvvqq8vDzl5uZq9+7d6unp0dq1a+0ezzH++te/SpK+//3v63vf+55+//vfa+3ataqtrdWpU6dsns45ksmkduzYobvvvluVlZV2j7NijIyM6NFHH9U3vvENu0cx1smTJ5VIJOT3++dt9/v9fMR9gebm5vStb31Ln/vc53T99dfbPY6xXBUnra2t8ng8530cO3ZMyWRS4XBYl19+uQ4dOqSBgQFt3bpVjY2N+sc//mH3j2G7xV7Hubk5SdJ9992nL3/5y6qoqNATTzwhj8ejZ555xuafwn6LvY6PPvqoTp8+rZ07d9o9spEWex3/29tvv60vfOELuv3229Xc3GzT5HCjcDis119/XU8//bTdoxjNVV9fPzk5qX/+85/n3eeqq67SoUOHdPPNN+udd96Z9yetP/7xj+uuu+5Sa2vrxR7VaIu9ji+//LJuuukmHTp0SJs3b069VlNTo2AwqPvvv/9ij2q0xV7Hr371q/rd734nj8eT2p5IJJSdna2vf/3r+vnPf36xRzXaYq9jTk6OJOnEiROqra3VjTfeqL179yory1X/Hy0ts7OzWrNmjX71q1/Nu8suFArp3Xff1f79++0bzoFaWlq0f/9+9fX16corr7R7HKOtsnuA5XTZZZfpsssu+9D93nvvPUla8D9aWVlZqXcD3Gyx17GiokJer1dvvvlmKk7OnDmj48ePa+PGjRd7TOMt9jo+8sgj2rVrV+r5iRMnVF9fr+7ubtXU1FzMER1hsddR+vc7JnV1dal38QiT88vJyVFFRYV6e3tTcTI3N6fe3l61tLTYO5yDJJNJ3XPPPfrNb36jF198kTBZBFfFyWJt2rRJa9euVSgUUltbm1avXq3HH39co6OjuvXWW+0ezzHy8/N19913q729XSUlJdq4caMeeughSdLtt99u83TOsWHDhnnPL7nkEknS1VdfrfXr19sxkiO9/fbbqq2t1caNG/Xwww9rcnIy9VpRUZGNk5ktEokoFAqpsrJS1dXV6uzsVDweV1NTk92jOUY4HNZTTz2l/fv3Ky8vL7Vex+fzafXq1TZPZybi5CwKCwvV09Oj++67TzfddJPOnDmj6667Tvv371cgELB7PEd56KGHtGrVKt1xxx3617/+pZqaGj3//PMsLMayO3jwoEZGRjQyMrIg6lz06Xbatm/frsnJSbW1tWl8fFzl5eXq6elZsEgW5/aTn/xEklRbWztv+xNPPKEdO3Ys/0AO4Ko1JwAAwHx84AoAAIxCnAAAAKMQJwAAwCjECQAAMApxAgAAjEKcAAAAoxAnAADAKMQJAAAwCnECAACMQpwAAACjECcAAMAoxAkAADDK/wMvRqWeJqCIjQAAAABJRU5ErkJggg==",
+      "text/plain": [
+       "<Figure size 640x480 with 1 Axes>"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    }
+   ],
+   "source": [
+    "charges = [np.sum(c.arrays[\"mbis_multipoles\"][:,0]) for c in good_configs]\n",
+    "\n",
+    "plt.hist(charges, bins=100)\n",
+    "plt.yscale(\"log\")\n",
+    "plt.show()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "mace-openmm",
+   "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": "3.11.9"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 2
+}

mace_dataset_prep.sh

@@ -0,0 +1,12 @@
+#!/bin/bash
+python ~/software/mace-tools/scripts/preprocess_data.py\
+    --train_file="/home/jhm/maceoff_scf_fitting/spice_v2_15_eV-A_force_filter.xyz"  \
+        --valid_fraction=0.05     \
+        --test_file="SPICE_v2_test.xyz"  \
+        --atomic_numbers="[1, 6, 7, 8, 9, 15, 16, 17, 35, 53]"  \
+        --atomic_multipoles_key="mbis_multipoles" \
+        --r_max=6.0 \
+        --h5_prefix="processed_maceoff_scf_spice_2/"   \
+        --compute_statistics    \
+		    --num_process 32 \
+        --E0s="{35: -70045.28385080204, 6: -1030.5671648271828, 17: -12522.649269035726, 9: -2715.318528602957, 1: -13.571964772646918, 53: -8102.524593409054, 7: -1486.3750255780376, 8: -2043.933693071156, 15: -9287.407133426237, 16: -10834.4844708122}" \

maceoff_1_with_multipoles_test.xyz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b459171ff8291b98e1675c72d03f3adec5df58aea5ecf74db63796175c0f10e8
+size 303730169

maceoff_v1_spice_force_filter_prep.sh

@@ -0,0 +1,12 @@
+#!/bin/bash
+python ~/software/mace-tools/scripts/preprocess_data.py\
+        --train_file="/home/jhm/maceoff_scf_fitting/maceoff_1_with_multipoles_train_force_filter_15.xyz"  \
+        --valid_fraction=0.05     \
+        --test_file="./maceoff_1_with_multipoles_test.xyz"  \
+        --atomic_numbers="[1, 6, 7, 8, 9, 15, 16, 17, 35, 53]"  \
+        --atomic_multipoles_key="mbis_multipoles" \
+        --r_max=6.0 \
+        --h5_prefix="processed_maceoff_scf_spice_1_low_force_filter/"   \
+        --compute_statistics    \
+		    --num_process 32 \
+        --E0s="{35: -70045.28385080204, 6: -1030.5671648271828, 17: -12522.649269035726, 9: -2715.318528602957, 1: -13.571964772646918, 53: -8102.524593409054, 7: -1486.3750255780376, 8: -2043.933693071156, 15: -9287.407133426237, 16: -10834.4844708122}" \

model_fitting_no_mixing/checkpoints/mace_pol_spice_v1_run-123_epoch-0.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:62141bcd6dba1d92040157bfdf66e881d03e75e232ca810da1bddffd4bccfe6d
+size 20837682

model_fitting_no_mixing/checkpoints/mace_pol_spice_v1_run-123_epoch-100.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:68909895723e7339abacf3669d0ac9befab47126032bc3493f2dea9a8dbbe010
+size 20838268

model_fitting_no_mixing/checkpoints/mace_pol_spice_v1_run-123_epoch-106.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e71f87f0cc9e7d1d1892df0cefca290cfbcc9259570320ed0549c96e77f983fb
+size 20838268

model_fitting_no_mixing/checkpoints/mace_pol_spice_v1_run-123_epoch-118.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9dccc9ab1a60bf7637a599fa20cddca75985275cd3b0027b04bbad2449c1c67d
+size 20838268

model_fitting_no_mixing/checkpoints/mace_pol_spice_v1_run-123_epoch-124.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:42efeb485e71144f7cd2f931ee085c74d2ca01cf9674f58f39a77f871dc386ce
+size 20838268

model_fitting_no_mixing/checkpoints/mace_pol_spice_v1_run-123_epoch-134.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:098829fec60d2ccdfbb95cfb285afa2286a82af80673cfee93c8db7e5c9cdbc7
+size 20838268

model_fitting_no_mixing/checkpoints/mace_pol_spice_v1_run-123_epoch-14.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a0e2198d09756cdd93cbcd89be36195f78c855395c1237541b1b7f0694b0f426
+size 20837975

model_fitting_no_mixing/checkpoints/mace_pol_spice_v1_run-123_epoch-140.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2cbca5e5b6b1b3b1b99375ac4aded5b52a3f4684be608c7449bba7c99a4862bc
+size 20838268

model_fitting_no_mixing/checkpoints/mace_pol_spice_v1_run-123_epoch-144.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cd713719731f8c9730b0ebdce02e2c08fb2ddddb74d26285d7dd672d0a09e432
+size 20838268

model_fitting_no_mixing/checkpoints/mace_pol_spice_v1_run-123_epoch-156.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9bac17cfa3c969e7688115bde9d79c433c55900b4f395305c3eed224154800b9
+size 20838268

model_fitting_no_mixing/checkpoints/mace_pol_spice_v1_run-123_epoch-160.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:21a4f224d14553dad138290cdbc491b9567c82fe1b0ce96f65a7130622d0d32e
+size 20838268

model_fitting_no_mixing/checkpoints/mace_pol_spice_v1_run-123_epoch-182.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:131299ffb110eb71582c891722bc3a516e6265526213dafa3c7e2104449988be
+size 20838268

model_fitting_no_mixing/checkpoints/mace_pol_spice_v1_run-123_epoch-196.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:12c21cf0152467129011f63ce1ae2da79dfa528172a3776fc5013833a9c501ec
+size 20838268

model_fitting_no_mixing/checkpoints/mace_pol_spice_v1_run-123_epoch-198.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a867b1fc85fd9611bde9e710ae50293b2c6865042069c7c868058af5aea3b93c
+size 20838652

model_fitting_no_mixing/checkpoints/mace_pol_spice_v1_run-123_epoch-2.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:be67841ab0d4c562d10d72c3c04721954d1ab28bd089f19e8225d29f23d58a19
+size 20837682

model_fitting_no_mixing/checkpoints/mace_pol_spice_v1_run-123_epoch-226.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:412db1477478c214a8fbb02bc693e6175e8d984caac47d1346e202bc49465a84
+size 20838652

model_fitting_no_mixing/checkpoints/mace_pol_spice_v1_run-123_epoch-242.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ad64712819e11416d2226be909260a6be935fde5549233a11908c9e63ea74246
+size 20838652

model_fitting_no_mixing/checkpoints/mace_pol_spice_v1_run-123_epoch-30.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:90fda6d8c8976781e3eec927999540807721e9b9d62955979b685b49fdac4337
+size 20837975

model_fitting_no_mixing/checkpoints/mace_pol_spice_v1_run-123_epoch-300.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1bc44af411552517f443ad2152d0eb7f5a0b650a0c45637faad6d7bd2d971a15
+size 20838652

model_fitting_no_mixing/checkpoints/mace_pol_spice_v1_run-123_epoch-314.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6fafe21458a82c3a28e1cf3a3949e1622f4de5f2193a3e668a3cd1f12c40206b
+size 20838652

model_fitting_no_mixing/checkpoints/mace_pol_spice_v1_run-123_epoch-318.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b00ed3abec63c93baa21385ac7670be145dcb0b4c6e7f1bfedb78a1ec70936b3
+size 20838652

model_fitting_no_mixing/checkpoints/mace_pol_spice_v1_run-123_epoch-32.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2fd2a9a9568ca8752152c236ef4e1516605c0ea77f3326d79bc42a0004fff0bb
+size 20837975

model_fitting_no_mixing/checkpoints/mace_pol_spice_v1_run-123_epoch-326.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:966076e3b9b011ca3c9534cf0e2ab37d6508fd1e6fe4681a14877bd4f7c825a6
+size 20838652

model_fitting_no_mixing/checkpoints/mace_pol_spice_v1_run-123_epoch-336.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9a406d0148d8065654bda1451ed930a6fde8ed5b5590ad71caa9c06f95c0825b
+size 20838652

model_fitting_no_mixing/checkpoints/mace_pol_spice_v1_run-123_epoch-360.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:97831f30a490ea3dcbe9b68771dd5d9ad9b1aea140014174082bcd38c46a2484
+size 20838652

model_fitting_no_mixing/checkpoints/mace_pol_spice_v1_run-123_epoch-370.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:858c6526d6f42b99fa2d5a9e89b692ff25d47ccaae6fd6576b2ed260a3f4a980
+size 20838652

model_fitting_no_mixing/checkpoints/mace_pol_spice_v1_run-123_epoch-382.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9852da2320f851f0c777cb12184086e9145d8ea80ae1442410a92a091e88669a
+size 20838652

model_fitting_no_mixing/checkpoints/mace_pol_spice_v1_run-123_epoch-40.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5d9df8cceba952a3d492d9b91f866e269d11d5a94af6c6193656728af0fe60a4
+size 20837975

model_fitting_no_mixing/checkpoints/mace_pol_spice_v1_run-123_epoch-402_swa.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:07169540b8c8f4a690fb2412a76ba1419de1e43338df4e3077b541f451d10b46
+size 20839952

model_fitting_no_mixing/checkpoints/mace_pol_spice_v1_run-123_epoch-404_swa.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:44546e2eafc8f4c54be3fa47686e6e5ac2aafd5a894fa7b35c41f93315577e11
+size 20839952

model_fitting_no_mixing/checkpoints/mace_pol_spice_v1_run-123_epoch-406_swa.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:60968ae8a1c89fab6db30393d91568011d6053c9f051ff5056a814dd2fb64d62
+size 20839952

model_fitting_no_mixing/checkpoints/mace_pol_spice_v1_run-123_epoch-408_swa.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:68be5c22a8cd6a2059afa0bb57259c90007b6f0a24eb0e565e82937c0b8e428f
+size 20839952

model_fitting_no_mixing/checkpoints/mace_pol_spice_v1_run-123_epoch-414_swa.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:346f18182ffc220e76ab69b016e7ae7d3aee6585f1f50694dc2b2b8dd27d1867
+size 20839952

model_fitting_no_mixing/checkpoints/mace_pol_spice_v1_run-123_epoch-418_swa.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a5441d421330712b1815c6373ba7b3d1e976257983a239288cb69375abddfeaa
+size 20839952

model_fitting_no_mixing/checkpoints/mace_pol_spice_v1_run-123_epoch-420_swa.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1d6da2b9c67f25ccf3a2a3b5629dd55bfd8e8cfdcd8f741ef21db1b14bd26ff7
+size 20839952

model_fitting_no_mixing/checkpoints/mace_pol_spice_v1_run-123_epoch-424_swa.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b528dc9ab5be86d6c4ac9026b3aeda004dc6891139864b383878bd1234912b9d
+size 20839952

model_fitting_no_mixing/checkpoints/mace_pol_spice_v1_run-123_epoch-426_swa.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:db40a9f85b27cb112a0e9f6bcdd18c02cdc64d783bfa8644dc387444ff9f3a39
+size 20839952

model_fitting_no_mixing/checkpoints/mace_pol_spice_v1_run-123_epoch-434_swa.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9276de71f4932cd33d8f5df897091223159ae2e05f85b24f405c0acc19075179
+size 20839952

model_fitting_no_mixing/checkpoints/mace_pol_spice_v1_run-123_epoch-436_swa.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3791a9a8a7549b6d37e07883e0b9c53336ac703244c5a7d84931e84021df82af
+size 20839952