{
"cells": [
{
"cell_type": "markdown",
"metadata": {
"nbsphinx": "hidden"
},
"source": [
"It looks like you might be running this notebook in Colab! If you want to enable GPU acceleration, ensure you select a GPU runtime in the top-right dropdown menu 🔥"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"# Fine-tuning\n",
"> **FYI**, you can open this documentation as a [Google Colab notebook](https://colab.research.google.com/github/jla-gardner/graph-pes/blob/main/docs/source/quickstart/fine-tuning.ipynb) to follow along interactively\n",
"\n",
"`graph-pes` provides access to several foundation models (with more being added as they are released) for you to use as-is, as well as to fine-tune on your own dataset.\n",
"See the [documentation](https://jla-gardner.github.io/graph-pes/interfaces/mace.html#mace-torch) for a growing list of the interfaces we provide.\n",
"\n",
"Below, we:\n",
"- install the relevant packages to use the [MACE-MP-0](https://github.com/ACEsuit/mace?tab=readme-ov-file#pretrained-foundation-models) and [Orb](https://github.com/orbital-materials/orb-models) families of models\n",
"- fine-tune the model on structures labelled *with a different level of theory*\n",
"\n",
"Since ``graph-pes`` provides a unified interface to many different foundation models, swapping between these different foundation models is incredibly simple!\n",
"\n",
"## Installation\n",
"\n",
"To make use of the `MACE-MP-0` and `Orb` families of models, we need to install `mace-torch` and `orb-models` alongside `graph-pes`:"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"!pip install graph-pes mace-torch orb-models"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Loading and using foundation models\n",
"\n",
"Let's start by loading the `MACE-MP-0-small` model and using it to make predictions on some SiO$_2$ structures. To do this, we'll use [load-atoms](https://jla-gardner.github.io/load-atoms/) to download the [SiO2-GAP-22](https://jla-gardner.github.io/load-atoms/datasets/SiO2-GAP-22.html) dataset."
]
},
{
"cell_type": "code",
"execution_count": 1,
"metadata": {},
"outputs": [
{
"data": {
"application/vnd.jupyter.widget-view+json": {
"model_id": "72f5c3fabe9546a9a3c8625fc32bf196",
"version_major": 2,
"version_minor": 0
},
"text/plain": [
"Output()"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"text/html": [
"\n",
"
\n",
"\n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
" \n",
"\n",
"\n",
" \n",
"\n",
"
"
]
},
"execution_count": 2,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"from load_atoms import view\n",
"\n",
"structure = dataset[0]\n",
"view(structure, show_bonds=True)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Now that we have a structure, lets use the [graph_pes.interfaces.mace_mp](https://jla-gardner.github.io/graph-pes/interfaces/mace.html#graph_pes.interfaces.mace_mp) function to load the `MACE-MP-0-small` model to make some predictions:"
]
},
{
"cell_type": "code",
"execution_count": 3,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Using device: cpu\n"
]
},
{
"data": {
"text/plain": [
"3847696"
]
},
"execution_count": 3,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"import torch\n",
"from graph_pes.interfaces import mace_mp\n",
"\n",
"device = \"cuda\" if torch.cuda.is_available() else \"cpu\"\n",
"print(f\"Using device: {device}\")\n",
"\n",
"mp0 = mace_mp(\"small\").eval().to(device)\n",
"sum(p.numel() for p in mp0.parameters())"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"This `model` object is a [GraphPESModel](https://jla-gardner.github.io/graph-pes/models/root.html) instance, and so can be used throughout the rest of the `graph-pes` ecosystem. For instance, we can inspect the dimer curves for this model:"
]
},
{
"cell_type": "code",
"execution_count": 4,
"metadata": {},
"outputs": [
{
"data": {
"image/png": "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",
"text/plain": [
""
]
},
"metadata": {
"image/png": {
"height": 309,
"width": 354
}
},
"output_type": "display_data"
}
],
"source": [
"from graph_pes.utils.analysis import dimer_curve\n",
"import matplotlib.pyplot as plt\n",
"\n",
"%config InlineBackend.figure_format = 'retina'\n",
"\n",
"for dimer, c in [(\"Si-Si\", \"crimson\"), (\"O-O\", \"green\"), (\"Si-O\", \"royalblue\")]: \n",
" dimer_curve(\n",
" mp0, dimer.replace(\"-\", \"\"), units=\"eV\", rmin=0.7, rmax=4.0, label=dimer, c=c\n",
" )\n",
"plt.legend();"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"... and also use it as an ASE calculator to generate force predictions:"
]
},
{
"cell_type": "code",
"execution_count": 5,
"metadata": {},
"outputs": [
{
"data": {
"image/png": "iVBORw0KGgoAAAANSUhEUgAAAgEAAAIBCAYAAADQ5mxhAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjkuMywgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/GU6VOAAAACXBIWXMAAB7CAAAewgFu0HU+AAB9JklEQVR4nO3dd3iTVf8G8PtJN02hg07KamnZlFVEhuxtZcoGQVBUfH+o8MorLhz4Ag5UfBURBEFlo+wpewiUjcjqkkKhG7pLk/P7oyZ2pBlt8qRt7s919bpCnnXS0ubOec75HkkIIUBEREQ2R2HtBhAREZF1MAQQERHZKIYAIiIiG8UQQEREZKMYAoiIiGwUQwAREZGNYgggIiKyUQwBRERENoohgIiIyEYxBBAREdkohgAiIiIbxRBARERkoxgCiIiIbBRDABERkY1iCCAiIrJRDAFEREQ2iiGAiIjIRtlbuwFkOQUFBbh37x4AwM/PD/b2/HETEdE/2BNQjd27dw9169ZF3bp1tWGAiIhIgyGAiIjIRjEEEBER2SiGACIiIhvFEEBERGSjGAKIiIhsFEOAFcyePRuSJGm/Dh06ZO0mERGRDWIIkNmFCxfw2WefWbsZREREDAFyUqvVeP7551FQUAAfHx9rN4eIiKqImzdvWuS8DAEy+vLLL3HmzBk0adIEU6ZMsXZziIioCti0aROaNWuGTz75xOznZgiQyV9//YW3334bALBkyRI4OjpauUVERFTZbdq0CaNGjUJBQQH+/e9/Y/fu3WY9P0OATKZPn47MzEw888wz6Natm7WbQ0RElZwmAKhUKgDA5MmT0bdvX7NegyFABuvXr8f27dvh6elpke4cIiKqXnQFgGXLlkGhMO/bNpeVs7D09HTMmDEDALBgwQLUrl3byi0iIiK5CSEQk5SF5Iw81HZzQkNvV0iSpHNfuQIAwBBgca+//jru3buHzp07m30wYHx8vN7tCQkJZr0eERGZ7mJcGr7edwtRiZna54J9lHipTyOE1fcotq+cAQBgCLCoo0ePYtmyZbC3t8eSJUvKTH3lVbduXbOej4iIzOtiXBpmr7mIArUo9nxUYiZmr7mIBWPCtEHg119/lTUAABwTYDH5+fl4/vnnIYTAq6++ihYtWli7SUREJCMhBL7ed6tUANAoUBduF6Jwe6NGjeDhURgI5AgAAHsCLOajjz7CtWvXUK9ePbz77rsWucbt27f1bk9ISECHDh0scm0iItIvJimr2C0AXaISMxGTlIUgHyVatGiBgwcPYsWKFfj4448tHgAAhgCLuHbtGv773/8CABYvXgxXV1eLXCcwMNAi5yUioopLzsgzar+UzDwE+SgBAC1atMCnn35qyWYVwxBgAYsWLUJ+fj6CgoKQnZ2NtWvXltrnypUr2scHDhzAvXv3AAAREREWCw1ERCSf2m5OercnXjmClBun4T75Z5laVBpDgAXk5RWmv+joaIwZM8bg/h988IH2cUxMDEMAEVEVVXQqoJfSEUE+rohOzCq1X+KVI7iy7n0ItRofzXkFy5Ytg52dneztZQggIiIyA11TAb3dnKCQgKJjA4sGAADaZeWtgbMDLGDlypUQQuj9KjpY8ODBg9rnGzRoYL2GExFRuWimApYcCJiUkac3AMg1C6AsDAFEREQVYGgqoEZlCwAAQwAREVGFGDMVsGQAmDRpktUDAMAQQEREVCGGpgKWDAD+bfvjzf9+YfUAADAEEBERVYi+qYBCrULsoR+LBYCmQ2chLfuRXM3TiyHASubOnasdDNi9e3drN4eIiMqpobcrgv8u9lOSpLBD60kL4OrbUBsAJIUCXkr9NQTkwhBARERUAZIk4aU+jWCv0D3Nz1HpgXbPfaENAMG+SjT0rhz1YBgCiIiIKiisvgcWjAlDsI8SqbfOoiAvu9h2BxclJIUC9goJ0/uEWK0uQEkMAURERGYQVt8DfdzjcGnVf3Bvy3vwraEutj3YR4mFY1ujVT136zRQB1YMJCIiMoNNmzZh9OjRUKlUuH7xDJ7NP4sRU19CSmYevJROaOjtWml6ADQYAoiIiIooWv+/tptxb96bNm3CqFGjoFKpABQWApo1axYUCoV2hcDKiCGAiIjob7rq/wf7KPFSn0YIq++h8xhdAaAyFAIyRuVvIRERkQzKqv8flZiJ2Wsu4mJcWqljqnIAABgCiIiIDNb/L1AXbhfin+1VPQAADAFERERG1f+PSsxETFIWAODIkSNVPgAADAFEREQG6/9rpGTmQQgB74bN0fGJngAqz2JA5cGBgUREZPP01f8vKuVhHl5YHomoxEw4dp+JUNcmcHh8HC7fflDmwMHKrOrFFiIiIjPTV/9fw9fNHot2XdfeNlDYO6Lu40MRnZxd5sDByo4hgIiIbJ6h+v/JV49g24fjkJl6T+d2XQMHqwKGACIiIgCt6rljxoBQ+Ls7F3teun0al9d+gIzE2zi3/DXkZ+r+xF904GBVwTEBRERk83QVCfKv5Yz6OVcw/7s3of57FoBHw9ZwqFGrzPOkZOZV6gqBJTEEEBGRTdMUCSpZI+Di8b34ad37EOrChYD82/bXLgdcFi+lcQMMKwveDiAiIptVVpGgxCtHcKVIAAjt/CR6TX5bbwAI9lWioberRdtrbpU+BMTExOCDDz5A48aNrd0UIiKqZnQVCSoZAPzb9kdg/1eQmJkPuzLWEbJXSJjeJ6TSrRJoSKW8HZCeno7169dj9erVOHHihLWbQ0RE1VTJIkG6AoDmFoBKLRDg4QxneztEFxkAGOyjxPS+IWhVz13OpptFpQkBBQUF2LFjB1avXo0dO3YgPz8fALTTLTw8ql4RBiIiqtxKFgl6GH9N7xiAu2m5WDqlPSBJSMnMg5fSuKWGKyurh4BTp05h9erVWLduHVJTUwH888bv5eWFwYMHY8SIEejVq5c1m0lERNWQpkiQ5pZAcL/nIIQKj7IzyhwEmJKVj/Agryo1C6AsVgkBMTEx+PHHH/Hjjz/i1q1bAP5545ckCYMHD8bLL7+M7t27w87OzhpNJCIiG6ApEqSZHSBJEhr1fwEQosxBgFVtBoA+ZgsBJ0+exIkTJ+Ds7Iy+ffsiJCSk2PYHDx5g3bp1xe7za9747e3t0bdvX+zcuRMAMGbMGH7yJyIii9u0aRP8/f2xYEwY/rf3JqKTsgq79svo3q+KMwD0qXAIePToEcaOHYvNmzdr39Tt7Ozw9ttvY86cOdi5c6f2Pn9eXuEADCEK01anTp0wbtw4jBw5El5eXlVyBSYiIqqaNm7ciNGjR6NGjRrYvXs3vp36OHZfSsDnu25ApS5d/reqzgDQp8Ih4K233sKmTZtQr149LF26FLVq1cKMGTPw3nvv4bPPPkNGRgaAfz71N2/eHOPGjcOYMWNQv379il6eiIjIZJoAoFKpkJGRgbVr16JTp04YEBaAAHeXUtUDq/IMAH0qHAJWrVoFSZLw6aefom/fvgCAH374AU2bNsXDhw8BAPXq1cOYMWMwduxYtGzZsqKXJCIiKreiAQAAJk+ejM8//1y7Pay+B5ZMaY+YpKxqMQNAnwqHAM0bfUBAgPY5f39/AIUDLlxdXfHMM89g/PjxpcYJEBERyUlXAFi2bFmp29GSJCHIR1ktZgDoU+Gb8J06dQIALF68WNvl/8UXXwAovAWQmZmJDz/8EE2aNEF4eDgWLVqEe/d0L8VIRERkKiEEohMzcToqBdGJmcWW8y267X/LfzQqANgSSVRw8ePLly+ja9euyMjIQN26deHq6oo///wTTZs2xdtvv401a9Zg9+7dePToUeEFJQkKhQLdunXDuHHjMHz4cNSsWRMAoFAoIEkS1qxZg5EjR1b81dm4+Ph41K1bFwBw+/ZtBAYGWrlFRETmpWv1v2AfJV7q0wgAtNvuXzmMP9Z9oC0ExABQqMIhAACio6Px4Ycf4tSpU3ByckLv3r3x5ptvolatwuUWU1NTsXbtWqxatQqnT58uvPDf91acnJwwcOBAjBkzBk8//TRDgBkxBBBRdVbW6n8ACmv8SxJUaoHc9Ps48dkECFUBAKBOu/7YtuFHtGnoJXOLKx+zhABT3Lx5E6tWrcJPP/2E2NjYwkb8HQg0UwcXLFiAWbNmydmsaokhgIiqKyEEXlgeWWrxn7LcPbsLf/7yCfzb9EPTobPQyK8mlkxpXy0H+5lC9hBQ1LFjx7B69Wps2LAB6enphQ36+wfi6+uLoUOHYvjw4ejevbvNd9mUB0MAEVVX0YmZeH7ZGZOOSYu9BPd6LbSVAJdODa/2A/8Mseo7a5cuXfDtt9/i3r172LBhAyIiImBvbw8hBO7du4clS5agT58+8PPzw/PPP2/NphIRUSVScvW/knIfJJV6zqNBq2KlgFMy9Z/DFlSKj9eOjo4YPnw4tmzZgrt37+LLL79Ehw4dIISAEALJyclYvny5tZtJRESVRMnV/4q6f+UwTnw6HgkX9uk9R3VaA6C8KkUIKMrLywsvv/wyfv/9d1y/fh1vvfUWKwsSEVExmtX/StLOAlA9wtWN85Eed0Xn8Y72CjSoXcPSzaz0LBYCLl68WOFzhISE4P3330dMTAwOHz5shlYREVF1IEkS+rXyK/ZcyWmA/m36oVbdZjqPzy9QIzY52+LtrOwsFgLatm2LBg0aYPr06di9ezfy8/MrdL4uXbqYqWVERFTVXYxLw9IDUdp/lwoAbfuj6dBZZS4HDHBMAGDBECCEwF9//YUlS5Zg0KBB8PLywtChQ/H999/j/v37lrosERFVc0IIfL3vlrY+QHkCAMAxAYAFQ0B8fDyWLFmCgQMHwsnJCVlZWdiyZQuee+451KlTB4899hg+/PBDs9w2ICIi2xGTlKWtD1DeABDsq0RDb1eLt7Wyk6VOQE5ODvbv34/t27dj+/btSEhIKLz43zUBAgMDMWjQIERERKBXr15wdHS0dJNsAusEEFF1dDoqBXPWXULilSO4su79MgOAQgJ0FBOEvULCwrGtq92ywOVhlWJBZ8+exbZt27B9+3acO3eusCF/B4IaNWqgd+/eiIiIwKBBg+Dr6yt386oNhgAiqmqEEIhJykJyRh5qu+lewldTKCgt9hIu/vAfqPJzdfYAzBrYGL9E3im1rsD0viEMAH+zasVAALh79y62b9+Obdu24cCBA8jJySlsmCRBkiS0a9cOERERiIiIQFhYmDWbWuUwBBBRVaJvMaCw+h7a59RqNZ785CjyC9RIi72ExEsHEfrkv4oFAEd7BbbP6gpJkhCTlIWUzDx4KXWHCltm9RBQVG5uLvbv36/tJSh526BOnTqIiIjACy+8gJYtW1qzqVUCQwARVRX6FgOyV0hYMCZMGwSMLRnMssCG2Vu7AUU5OzvjySefxJNPPgmg8LaBppfg3Llz2sGGvr6+DAFERNVEydH+JRWoC7f3rhWLixcvImLSK0adNyUzjyHAgEoVAkpq164d2rVrh3fffRd3797Ftm3bsGPHDtSowSpPRETVRdHR/mU5cWAHlq//ECqVCikPcyB8Bxns1ucUQMMsNkXw1VdfxYULF8x2voCAAEybNg1bt27lMsNERNXImagUvds10wBVKhUAIDcjFcEGpvdxCqBxLBYCvvjiC7Rr1w6tWrXCJ598or2/T0REpHExLg3fH4ouc3vJOgCTJ0/G8uXLMb1fKOwVunsC7BUSpvcJ4QBAI1hsYKDi71Gamh+CQqFAz5498cwzz2Do0KFwcXGxxGWpCA4MJKLKSgiBmMRMvLPhMu491F2+t2QAmDRpEpYvX659fylrNgGnABrPYiFg3759WL16NX755RdkZWUVXuzvQKBUKjF8+HBMmDABPXr0sMTlCQwBRFQ56XrzLqlkABj89DhsXrtKGwA0NHUFOAWwfCw+RTA7OxubN2/G6tWr8dtvv0H99w+0aLXA8ePHY8KECWjSpIklm2JzGAKIqLLRNxVQo2QA6D7oafy2dW2pAEAVJ2udgISEBPz000/48ccfcenSpX8a8XcgaNeuHZ555hmMHj0aXl5ecjWr2mIIIKLKRAiBF5ZH6u0BUBfk4/cvJiMntXAcmX/b/jiyYx0a+dWUq5k2xWrFgi5fvoxVq1ZhzZo1uHv3bmFj/g4DDg4O6N+/PyZOnIiIiAg4ODhYo4lVHkMAEVUmxhb5yUm9i3PLX4NHUFsMmvYulj73GLv4LcTqFQOFEPjtt9+wevVqbN68udT4AQ8PD4waNQoTJkxAx44drdnUKochgIgqE83CP8bIy0iFo6s7/h3RDP3D/C3cMttl9RsskiShd+/e+OGHH3D//n2sWrUKffr0gUKhgBACqamp+Oabb9ClSxdrN5WIiCqgtpvu4j1psZegLnhU7DknN09ICgW83LiqrCVZPQQUVaNGDYwfPx579uzBhQsX0Lx5c22PQCVa4oCIiAwQQiA6MROno1IQdT8D0fczkPwwF4GexaeH379yGOeXv4bLa98vFQQAVv2ztEpVNvjRo0fYtm0bfvzxR+zcuROPHpX+D0FERJWbMVMAgeKzAJL/PI67kTsR2HGwdrujvQINarNMvCVVihBw/PhxrF69Ghs2bEB6ejqAfz75u7m5YcSIEXjmmWes2EIiIjKGMVMAgdLTAP3b9kedDhHF9skvUCM2OZuLAFmQ1ULAzZs3sXr1avz000+IjY0F8M8bv52dHXr37o2JEydi6NChcHZ2tlYziYjISIZWA9TQFQCaDp0FSUcdAK4EaFmyhoDk5GSsXbsWq1evRmRkJIDi9/pbtmyJiRMnYty4cfDz85OzaUREVEHGrAZoSgAAOCbA0iweAvLy8rBlyxb8+OOP2LNnDwoKCgD88+bv6+uLsWPHYuLEiQgLC7N0c4iIyEKSM3SvAaBhagDgSoCWZ7EQcOjQIfz444/YtGkTHj58COCfN35nZ2c89dRTmDhxIvr16wc7OztLNYOIiGRS1hRAAEi5FWlSAOBKgPKwWAjo2bMnJEnSvvFLkoQuXbpg4sSJGDlyJGrWZAlIIqLqpKG3K4J9lDpvCdQKbAq3Ok3w8PZVwz0AXAlQNha9HSCEQHBwMCZMmIAJEyagYcOGlrwcERFZkSRJ6NfKD1/vv1Vqm72zK9pMWoD4U1tQv+voUgHgnWHN4eJox5UAZWaxEPD8889j4sSJ6NSpk6UuQURElciFuDR8eyBK+2+hVkFS/HO7197ZFQ26jS11XLCvEl0be/ON3wosVjFwyZIlDABERDbiQmwq5qy7BNXf0wPvXzmMM99MR37WA73H8d6/dVm1WFBUVBROnjyJe/fuITs7Gy+99BJq165tzSYREVERQgjEJGUhOSMPtd2ctKP1iz73MDsfs9dchOrvGd9FZwGc/34W2j23CPbOpef6B3q64LWBTXjv34qsEgLOnTuHV155BcePHy/2/IgRI4qFgP/973947733UKtWLVy9epVLChMRyUhX+d8AdxdAEriblqt9ztFeoTMAAIBbQAjsHEuX/g30dMH3z3eAoozBgSQP2b/727dvR+fOnXH8+HEIIbRfukycOBE5OTmIjo7G9u3bZW4pEZHt0pT/LTnS/256TrEAABSW9wWMrwNgr5Dw2sAmDACVgKw/gYSEBIwZMwZ5eXlo1qwZdu3ahYyMjDL3d3Nzw1NPPQUA2LVrl1zNJCKyacaW/y3K2AAQ5OOKhWNb8xZAJSHr7YBFixYhKysL9evXx9GjR+Hu7m7wmO7du2PNmjU4e/as5RtoZpGRkdi5cyeOHTuGq1evIikpCQ4ODggICEDnzp0xZcoUdOnSxdrNJCIqxpjyv0UZEwAUEvDawCbo18qPgwArEVlDwO7duyFJEmbOnGlUAACAJk2aAABiYmIs2DLze+KJJ3D06NFSz+fn5+PmzZu4efMmVq5ciYkTJ+K7776Do6OjFVpJRFSaofK/RRkTABztFfjvqFYIq+9h9rZSxcgaAuLi4gAAHTp0MPoYTWXBzEzjU2llcPfuXQBAQEAAnn76aXTt2hX16tWDSqXCyZMn8emnn+LOnTtYtWoVHj16hJ9//tnKLSYiKqSv/G9JKdd/1xsA7CQwAFRisoYAzeJB6r//wxjjwYPCOaZKZdVaSrJJkyb46KOPMHz48FJrI3Ts2BETJkxA586dcePGDaxZswYvvPACnnjiCSu1lojoHw9zHhm9b9OhsyBUKkh29mg5/N9Q4Z+ufpb/rfxkDQF+fn6IjY1FdHQ0OnbsaNQxp0+fBgDUq1fPkk0zO0OzGWrXro1PP/0UERERAICNGzcyBBCR1Qkh8M2+0mV/yyIp7NBsxGwAEl4d2AShATWRkpnH8r9VhKyzA7p27QohBDZs2GDU/vn5+fj2228hSRK6d+9u2cZZQY8ePbSPo6Ki9OxJRCQPQ4MCk64eQ2ZibLHnJIUdJIUCfu7OCPJRIjzIC0E+SgaAKkDWEDBp0iQAwNatW7Fv3z69++bn52PixImIioqCJEl47rnnZGihvPLy/hl8w+WUiagy0Dco8P6Vw7i8Zi7OL59ZKggAwDf7o8qs+0KVk6whoHv37hg1ahSEEIiIiMDs2bO13f0AEBsbixMnTuDjjz9G8+bNsWHDBkiShBdeeAHNmzeXs6myOHz4sPZx06ZNrdgSIqJCZQ0KLDoLID8zDQlnS9duiUrMRExSlqWbSGYkCZljW15eHoYPH46dO3fq7SrSNGvYsGFYt25dtfukrFar8fjjj2tDUGRkJNq1a2fSOeLj4/VuT0hI0M7EuH37NgIDA8vXWCKyGUIIvLA8stgtAWMLAQHAf0e3QniQl2ztpYqRfe0AJycnbN++Hd999x0WLlxY5r3wwMBAzJkzBy+88ILMLZTHokWLtAFg2LBhJgcAAKhbt665m0VENk6SJLzYpxFm/XQBgGkBAAC8lMZPLyTrk70noKSrV68iMjISiYmJUKlU8PLyQps2bdC2bdtqO6jk8OHD6N27NwoKCuDj44PLly/Dx8fH5POY8v1hTwARGSs6MRPPLztjcgAI9lViybPtq+3f7urIqksJA0CzZs3QrFkzazdDNn/88QeGDh2KgoICODs7Y8OGDeUKAEDhG7s+RW8HEBHpomup4OSMPJMDgL1CwvQ+IQwAVYzVQ4AtiYmJQd++fZGWlgY7OzusXbu2QrUB+MmeiCriQmwqPt99A/GpOdrngn2UaO+dY1oPAIsCVVkMATK5e/cuevfujbt370KSJHz//fcYPHiwtZtFRDZq8+nb+Hp/6aJAUYmZiEuWEPbk87iwdUmZAcC7phNeHdBY23vAHoCqiSFABsnJyejTpw+io6MBAIsXL8bEiROt3CoislUX4tJ0BgCNArVASM9xkNzrwSOkQ6kAYCcBbzzVjJ/8qwFZ6wTYogcPHqBfv364evUqAGD+/PmYPn26lVtFRLZKCIHPd10v9Xx+1oNi/07NeoThQyPQyK9mseeDfZT4eFwbBoBqgj0BFpSdnY1Bgwbh3LlzAIA333wTs2fPtnKriMiWxSRlFRsDABROA/xz0wK0HDMXXqH/DCY+eDUJn4wNQ80ajlwPoJpiT4CF5OfnY+jQoTh+/DgAYMaMGfjwww+t3CoisnUlywJrZgGo8nNx6ae3kZFQvHbL57tvoKG3K9cDqKbYE2AhY8aMwd69ewEAPXv2xJQpU3DlypUy93d0dERoaKhczSMiG5VaJASUnAbo26oXlL4Ni+0fn5qDmKQsBPlUreXcyTgMARayefNm7eMDBw6gVatWevevX78+YmNjLdwqIqrOdM35L/rJXQiBzWcKy42bUgcgJTOPIaCaslgIEEKw24iISCYX49Lw9b5bxWr+B/so8WLvYNSs4YjkjDzkPlIhOimLpYBJy2IhwMfHBwMHDsSTTz6J/v37w83NzVKXqpS4nCYRyeViXBpmr7mIAnXxvztRiZmY9fPFYs+VpxRwQ29XyzScrM5iAwNTUlLw448/YvTo0fD29kbv3r3xxRdfaOfKExFRxQghEHU/A5/suFYqAOhiagCwU0h4qXcj9upWYxYLAT/99BNGjRqFWrVqIT8/HwcOHMBrr72GkJAQNG/eHLNnz8axY8f4iZmIqBwuxqXhheWRmLY8EgnpuUYdo1DYASh8QzcUAABApRb4et8tXIxLM0eTqRKy+CqCKpUKR48exbZt27B9+3bcvHmz8MJ/J0tPT08MGDBAe9ugZs2a+k5HJoiPj9cuN8xVBImqrpID/h5m5+M/ay8Z9em/pKSrx5B84zSaPPWK3gBQlL1CwoIxYQir72Hy9ahyk30p4Rs3bmDbtm3Ytm0bTpw4gYKCgsKGSBLs7e3RtWtXREREICIiAkFBQXI2rdphCCCq+nQN+HO0VyC/QG3W6xg6Z7CPEkumcJng6kb2EFBUeno6du3ahW3btmH37t1IT08vbNTf/8kaN26sDQSdO3fmfz4TMQQQVW1lDfgz1v0rh5GfkYK6jw/Tu9+4zvXx0/E4g+dbOjWcUwWrGavWCXB3d8eYMWMwZswYqFQqHDt2TNtLcPPmTVy7dg3Xr1/HJ598Ak9PT/Tv3x8REREYMGCAzc02ICLbIkTh/fiKBADtIEAB1O1UdhDwUjoadU7WC6h+Kk3ZYDs7O3Tr1g2ffPIJrl+/jmvXruHjjz/GE088ATs7O6SkpOCnn37CmDFjsGjRIms3l4jIomKSsordAjBFyVkAGfeiyhyEHeyrRPPAWkadl/UCqp9KEwJKCg0NxcyZM3Hw4EEkJiYWm21ARFTdlazxbyyd0wCHzNR5O9VeIWF6nxAE+SgRbOATPusFVE9VomxwydsGqamp1m4SEZFF1XYz/VO3KXUAgrxd8XK/UO2SwC/1aVTm+ANNWOC4rOqn0vYElMXOzg7e3t7WbgYRkUU19HY1+Om8KGMDgJ1CwqxBTfDt1HBtAACAsPoeWDAmrNQ1g32UWDi2dbF9qfqw6uwAsizODiCq2i7GpeH1ny9AZeCvtLEBINhHiel9Q/S+oWtqEqRk5sFLWXoRIqpeqsTtACKi6k7XCoCaT+f/WXcJBWUkgYK8HNzY9qXOABDk7YphHerCy83R6Dd0SZIQ5KPkLAAbwRBARGRlZa0A+FKfRgir7wGfmk64m6a7NLC9kwtaP7MA576fBe+mnbQB4J1hzdG1sTc/xZNeDAFERFakbwXA2Wsu4pX+oWUGAA23gEbo8NI3cHb3haRQINhXyQBARqlyAwOJiKoLQwWBCtRCZyW/h/HXtN3/Gi6e/pAUCo7kJ5OwJ4CISEZF7/3nPlIZLAiU8KB4L4BmEKB/2wFoMvjVYoP/NAP/WtathejEzGLjCxgKSBeGACIimei6928MD1cHpGU9KjYL4G7kDrg3aAn/Nn3h7+6MucNbIMhHiUt/peOF5ZFlji8gKoq3A4iIZKC591+eUsAlAwBQOAvAL6w37BUS/v1kUwT7uuHSX+k6r6EZX3AxLs0sr4WqD6v2BOTk5ODs2bO4d+8esrOzMWTIENSsWdOaTSIiMjuzLgaEf6YBNvKrqZ33b8z4gq/33eJywFSMVULA7du3MWfOHGzYsAGPHj3SPt++fXs0a9ZM++/ly5fj22+/Ra1atbB3717+xyWiKsmciwH5t+2PnpPfwntPt0KQj1L7d9GYa0QlZiImKYs1AEhL9tsBp06dQps2bfDzzz8jPz8fQogyV7eKiIjApUuXcODAAezdu1fmlhIRmYdZFwMaOgv3HuZDkqRiH4yMvUZKZvnaQtWTrCEgPT0dgwcPRmpqKvz8/PD111/j8uXLZe7v4+ODAQMGAAB27NghVzOJiMyqPIsBJV/7XW8p4JJv5sZeg8sBU1Gy3g748ssvkZiYiNq1a+PkyZOoV6+ewWN69+6NLVu24PTp0zK0kIjI/DSLAZlyS8CtTihcvAKRnfSXzrUAvFwdi00DbFC7hsFrcDlgKknWELBt2zZIkoTXXnvNqAAAAM2bNwcAREVFWbJpREQWI0mS3qV6dXFy80TbKZ/hzqmtaNhzYrEAEODhgvlb/0R0Upb2uWAfJfq18sPSA1FcDpiMJuvtgFu3bgEAnnjiCaOP8fAonNf68OFDi7SJiEgOZS3VW1TJ8VFObp4I6j2pWACwU0i4n55TLAAAhYP+lh6IwvM9g7kcMBlN1p6A3NzCylcODg5GH5OVVfgf3cXFxSJtIiKSS1h9DyyZ0h4xSVk4cTMZKw/HaLfdv3IYd89sR6tx78POUfffu0BPF2TnqZCala9ze4FaYM+le/jm2XaITc7mcsBkkKw9AT4+PgCAmJgYA3v+48KFCwCAgIAASzSJiMhoQghEJ2bidFQKohMzy5zZpI9mqd5OIbW1z2lmAaTeOosLq96E6tE/g/7eGdYcz/UIgm9NJ8Sn5pQZADSiEjMRm5yNIB8lwoO8ik0jJCpJ1p6Axx57DPHx8di1axdGjhxpcH8hBL777jtIkoSuXbvK0EIiIt0MLfdrKs1gwRMHdhSbBeDi4QeFXWFvabCvEkkPcvHdwWiTzp2SmcdaAGQUWXsCxo0bByEEfvrpJ+0nfH1mzpyJixcvAgCeeeYZC7eOiEi3skr+GirHq6/nQJIkhDz6o8xpgPYKCe0auuOb30wfFM1pgGQsWXsCBg8ejB49euDgwYPo1asXPvzwQwwfPly7vaCgAHfv3sXx48fx5Zdf4sSJE5AkCcOGDUOnTp3kbCoREQDjlvvVVY73Ylwa/rfvJqIT/xnAF+Tjiul9QhBW3wMbN27Ef/41VWcACPBwhlAD63+PN7m9nAZIppBEeW5qVUB6ejp69eqF8+fPG7xPJYRAx44dsW/fPri68j+1qeLj41G3bl0AhaWaAwMDrdwioqonOjETzy87Y3C/pVPDtV3wF+PS8Pqai1DpCA52Cgl9asXiP/+aCpVKBQCYNGkS3vzvF0jLfoSUh3lYtOs6VOX4y2yvkDgLgEwie9lgd3d3nDx5Em+88QZq1qypLRtc8svFxQWvv/46Dh06xABARFZjajleIQQ+3XlNZwAAgLuXDuH1l/8JAJMnT8by5cvRyK8m2jf0xC+Rd8oVADgNkMrDKgsIOTo6Yt68eZgzZw4OHz6MyMhIJCYmQqVSwcvLC23atEHv3r1Rq1YtazSPiEjL2HK8CWk5hff91WrcTcvVuY8QAgnn9kCo/wkAy5Ytg+LvOgDlWWjI280JH45syVkAVC6y3w4g+fB2AFHFCSHwwvJIo9+caysdkZxZ9jQ+1aN8XPrpbTQOqodjO9drA4AQAtvO3cGXe26a1L5Px7Uu1+wEIsBKPQFERFWFqSV/9QUAALBzcETY+A/gXauG9pP7hdhUfL77BuJTc0xqm2aQIVF5yR4C/vrrLwCAr68vnJz0d7Pl5uYiMTERAIxea4CIqDyEEIhJzMSV+IcABJoH1tJ2sWtK/pasE2CMpD9PwC0gBM61vLXPKewdkZJVgJikLJyPTcM3+2+ZdM5ATxe8OqAxAwBVmKwhYO/evRgwYACUSiViY2MNhoDs7Gw0b94cOTk52L9/P7p37y5PQ4nIplyMS8OnO6+Vupcf4O6CmYMK32yLlvz9I/4Bvth9w+B5NZUAnd390HbqomJBAAC2RMZjx4UEo9upkIDXBjZBv1Z+vP9PZiHr7IANGzZACIEhQ4ZoFwbSx9PTE8OHD4darca6detkaCER2ZqLcWn4988XdA7mu5ueg9d/vqAtBqQp+etby9ngeTUBQKjVyEm9izunt5bax5QAEOyjxCfj2qB/mD8DAJmNrCHg5MmTkCQJffv2NfqYfv36aY8lIjIntVqN+VuvQt+tfpUA/rf3ZrFqf4ZmDBQNAEBhIaCgXpPL3c4Z/UOxZEp7Tv8js5M1BMTGxgIAQkNDjT6mUaNGAExbdIiIyJCLcWkY97+TSMrQP5APAKKTshBTZOleTd1/XXQFAE0lwPJqHliLn/7JImQdE1BQUAAAsLOzM/oYzb6aZYiJiCrqYlwaXv/5gklFef6IT0fyw1zkFqjhZK/A4Hb++GxX8el8lggAgZ4uLANMFiNrCKhduzYSEhIQHR2Ntm3bGnVMdHTh6lmenp6WbBoR2QjNWgCmVuX7Yrf++fuWCAAAMKpjPfYCkMXIejugdevWAGDSIL+1a9cCAFq0aGGJJhGRjSlPVT5DMu7etEgAAIDaNbkiIFmOrCFg8ODBEEJg8+bN2LBhg8H9169fj82bN0OSJAwZMsTyDSSiau9MVIrZz6n0b4Q6jw0GYN4AAHBZYLIsWcsG5+fno0mTJoiNjYWdnR1mzJiBGTNmaEvbaty+fRuLFi3C4sWLoVarUbduXVy/ft1gXQEqjmWDiQoJIRCTlIUzUSlYfjAaagtd4/6lA/Bt2UNvAPBwdUBa1iOjzhnsq8SSZ9vzdgBZjOxrB1y4cAFPPPEEMjMztf+x69WrB39/fwBAQkKCtqqgEAJKpRKHDx9GmzZt5GxmtcAQQFQ4CPB/e28iusjofnMoyM2CvbPlBuwpJOCTcW04LZAsSvalhFu3bo1Tp06hTZs22mWD4+LicOrUKZw6dQpxcXHa59u1a4fTp08zABBRuWhmAZg7ANy/chgnPh2HB7f/NOt5i3ptYBMGALI4qywg1LRpU5w9exb79u3D9u3bcf78eSQnJwMonEHQtm1bREREoFevXtZoHhFVA0IIfLrjusmzAAwpOgvgwsrXEf7SEtTwqmPWawT7KtGvlZ9Zz0mki6whQNPNr1Qq4enpiT59+qBPnz5yNoGIbIAQAkeuJeJuummr8hlSchqgd7OucPHwN+s17BUSpvcJ4TgAkoWstwMaNGiAhg0baqf9ERGZ28W4NLywPBIf/HLVrOe1VB2AooJ9lFg4tjVvA5BsZO0JcHFxQW5uLsLDw+W8LBHZiPJUAjSGKQHA390ZLg52Ro9DeGdYc7g42sFL6YSG3q7sASBZyRoC6tSpg6ioKKhUKjkvS0Q2QI4xAID+AKAA8O8nm0IIgdlrLqJA38pEKLz337WxN9/4yWpkvR2gWT3w2LFjcl6WiGzA7ksJFh8DoC8ABLi74JPxhVP6wup7YMGYMAR4lL3kMO/9U2Uga52Amzdvok2bNlAqlTh79izq1DHviFoqjnUCyFao1WqM/d9JJBuxIqAp7pzejmtbPgOgOwAoJGBoeB30bemPIB9lqTd0IQT2XEzAT8fjkPDgn0XQgn2UmN43hPf+yepkLxa0detWjB8/HrVq1cKCBQswYsQIODo6ytkEm8EQQLbgYlwaFu26jvhU43sBgnxckZKRhwc5BQb3vXN6Ox7GX0OTIa/p7AEI9lFiyRT9Vf00FQtTMvN4758qFVlDQM+ePQEAcXFxiImJgSRJcHR0REhICDw8PPQuMSxJEn777Te5mlotMARQdXcxLs2oe+8a7jXs0bWxDw7+cR+Z+caPTRJC6H3TXjo1HEE+SqPPR1RZyDow8NChQ8V+kYQQyMvLw5UrV8o8RpIkg7+ARGQ7NJ+qkx/m4uv9t4wOAADwILsA287fLXP7/SuHAbUavq16FHve0N+flMw8hgCqkmQNAU888QTfzImo3C7EpeFzE7v+i9IXF7SDAP/eqWQQ0Icr/VFVJXtPABFReWw+fRtf779lkXOXnAWQGnXO6BAQ7KtEQ2/LLSREZEmyLyBERGSqC7GpsgUA/7b90WTwq0Ydy2l+VNVZZQEhIiJjCSHw+e4bFjl3RUoBc5ofVQdWDwFCCERHRyM1NRUA4OnpiaCgICZrIgIAxCRllXsMgD6mBoAgH1cMC68LLzdHTvOjasNqIWDPnj346quvcOjQIWRnZxfbVqNGDfTo0QMvv/yytsogEdkWzSyAEzeSTT420NNFb3AwJgA42En46pm2SM1+xDd9qrZkLxaUn5+PSZMmYd26dQAKf9F1NuzvX7ZRo0Zh5cqVLChUDqwTQFXVxbg0fL3vFqISM40+xq+WM/6vfwhquzkj6WEu3lx/Wed+j7If4vgn46DKK1zgp6wegOl9QjA0nL8zVL3J3hMwduxY/PLLLxBCwN7eHn369MFjjz0GPz8/AMC9e/dw+vRp7Nu3D48ePcK6detQUFCA9evXy91UIrICUwsAafz7ySYIq++BC7Gp+GT7tTL3c6hRE63Gv4+Lq+bAt2WPUgHA0V6B53oEMwCQTZC1J2DHjh2IiIiAJEno3r07vv/+e9SvX1/nvn/99ReeffZZHDhwAJIkYdu2bRg4cKBcTa0W2BNAVY0QAi8sjzSpBwAAXurdCGH13fHTsVgcuW7c7YPM+zFw9a6vDQAKCRjfpQHGd64PhREDA4mqA1n/p69cuRIAEBYWht27d5cZAACgXr162LVrF1q3bg0AWLFihQwtJCJriknKMikABHq64MXewfg18g6mLY8sMwBk3o8pdetR6duwWA+AWgDHryfzvj/ZFFlDwO+//w5JkjBz5kw4ODgY3N/BwQGzZs2CEAK///67DC0kImtKzsgzaf+HOY+wZH+U3iWE7185jNNfPYeY334ocwySRlRiJmKSskxqA1FVJmsISEpKAgA0a9bM6GOaNGkCAEhONn2EcGUSFxeHmTNnokmTJnB1dYWnpyfCw8Px8ccfl5odQWSraruZVn73YU6BcaWA1WrEHFyF5D+PGzxnSqZpQYSoKpN1YKCrqyvS09ORkpJi9DFpaWkACqcNVlXbtm3D+PHj8fDhQ+1z2dnZiIyMRGRkJJYtW4YdO3agUaNGVmwlkfU19HZFsI/S5DEBuuiaBli7SSeDx3EdALIlsvYENG7cGAC00wONodlXc2xVc/78eYwaNQoPHz6EUqnEvHnzcOLECfz222947rnnAAA3btzAoEGDkJGRYeXWElmXJEno18qvwucpbyVArgNAtkbWEPDUU09BCIEVK1ZoBwnqs3r1anz//feQJAlDhgyxePssYcaMGcjJyYG9vT327t2LOXPm4PHHH0fPnj2xdOlSLFy4EEBhEPj000+t3Foi6xJCYM+lexU6R3kDANcBIFsk6xTBrKwshIaG4t69wl/yAQMG4Nlnn8Vjjz0GHx8fSJKE+/fv49SpU/j++++xa9cuCCFQp04dXL9+vcrdEjh9+jQee+wxAMC0adOwZMmSUvuo1Wq0aNECf/75J9zd3ZGYmGjUoEljcIogVTXRiZl4ftmZch9vbABwtFcgv0Ct/TfXASBbJfuYgO3bt6N3795IS0vDrl27sGvXrjL3F0LAw8MD27dvr3IBAAB+/fVX7ePJkyfr3EehUGDixIl44403kJ6ejoMHD7JUMtksU2cHFJV09bjRPQD5BWq8M6w5XBztWBKYbJrsFTHatGmDy5cvY/jw4VAoFBBC6PxSKBQYMWIELl26hLCwMLmbaRbHjh0DUBh+2rVrV+Z+3bp10z4+ftzw6GWi6srU2QFFufrUh6PSE4BxtwBcHO0QHuSFIB8lAwDZLKssIBQQEIANGzYgISEBhw4dwpUrV4qtItiiRQt0794d/v7+1mie2fz5558AgEaNGsHevuxvtWYaZNFjiGxRRWYH1KgdiLZTPsOdyB1o1Pc5g2MAOAuAyEIh4MsvvwQATJgwAR4eHmXu5+/vjzFjxliiCVaXm5urrW1g6F68h4cHXF1dkZWVhdu3bxt9jfj4eL3bExISjD4XUWWgmR3w9f5b5Tq+Ru1AhPSfZnA/zgIgKmSREPDKK69AkiT07t27WAjo2bMnJEnSu2ZAdVF0up9SqTS4vyYEZGYa/wlIM+iPqLoQQmD3JePC6/0rh5H0x1E0G/EfKOyM/1PGWQBE/5D1dsChQ4cgSRKysqp/Wc7c3FztY2OWQXZyKuyazMkpu/wpUXW352ICohMN/30oOgtAqAvQfORbRgUBBzsJ80eHcRYA0d8sEgKcnZ2Rl5eH9PR0S5y+SnB2dtY+zs/PN7h/Xl7hqGgXFxejr2Ho1kFCQgI6dOhg9PmIrOliXBo+3Xnd4H4lpwHaOdaAJBk3xvmRSsDNxTxTcImqA4uEgAYNGuD69evYvn07OnUyXKazOnJzc9M+NqaLX9M7YsytAw3O+6fqQgiBT3dc17sOAFD+QkBFpWTmIcjH+N8zourMIiFg4MCBuHbtGhYsWIDffvsNoaGhxQrgvPXWW3B3dzfpnJIkYfny5WZuqeU4OzvDy8sLKSkpBgfwpaWlaUMA7/OTLYpOzNS7EiBgngAAcFYAUVEWCQFz5szB1q1bcevWLZw5cwaRkZHabUIIbNmyxaTzCSGqXAgACldLPHr0KG7duoWCgoIypwleu3ZN+7hp06ZyNY9IVkIIxCRlITkjD7Xd/inQI4TA4T8T9R5rrgDAWQFExVkkBHh6eiIyMhJfffUVfvvtN9y5cwd5eXmIi4uDJEnw9/c3W2ncyqxLly44evQosrKycPbsWW0J4ZIOHz6sfdy5c2e5mkckm4txafh6361i8/+DfZTo18oPey7d01sXwFwBgLMCiEqTde0AhUIBSZJw+fJlNGvWTK7LWg3XDiAqDACz11xEgdr0PzVCrcbZ7/4PD/66CqACPQBcG4BIJ6tUDLQVHTp0QNeuXXH06FEsX74czzzzDB5//PFi+3z66afaKoEzZsywiR4Ssh1CCHy971a5AgAASAoFwibOx/kV/4bSt6FJAeCdoc3g4mTPtQGI9JC1J0DT7d2hQweTpsJVZefPn0fnzp2Rk5MDpVKJOXPmoEePHsjJycHatWuxdOlSAEBoaCgiIyOLzSqoKPYEkLVVdFVAjYK8bNg5OBsdAAI8XPDDC4/xjZ/IAFl7AooulGMr2rRpg3Xr1mH8+PF4+PAh5syZU2qf0NBQ7Nixw6wBgKgyKM+qgCk3z6BmYBM4uPzz+2DvZPwqogoJmDWoCQMAkRFkX0XQFkVERODSpUt49dVXERoaiho1asDd3R3t27fHggULcP78eTRq1MjazSQyu5RM00LA/SuHcXHVGzj//Sw8yskwfIAOU3sE8d4/kZFkvR1A8uLtALImIQSmLT9jVBlgoPQsgKDez6Jhj/EmX3fp1HAWAyIyEnsCiMgiYpKyyh0A/Nv2R4NuY02+JusAEJmGIYCILMLY8QDmqgNgJ4F1AIhMxCmCRFQhZVUCrO1muDyvuQIAAHgqncC7m0SmYQggonIrqxLgS30aoVU9dwT7KMusBmjOAAAASRl5mL3mIhaMCUNYfY9ynYPI1vB2ABGVi6YSYMk3+ajETMxecxGX/krHi72DdR6bHnfFrAFAo0BdWJyIPQJExmEIICKTGaoEWKAWWLTrOv5Kzta5vWZgU3g3fwKA+QKARlRiJmKSjBuQSGTreDuAiEwWk5Sld9EfAIhPzcGXe2/q3Kaws0Pzp9+EZ1BbBLQfaLYAoJGSmcdpgkRGYE8AEZmsPJUAVY/yi/1bYWeHOh2eNCoAeLs5wMPV+HU1vJSGByUSkYV6Anr27Gn2c0qShN9++83s5yUi0xkz8r+o+1cO49aub9Bm8seoUbuuyddLynhk9L6sFUBkPIuEgEOHDkGSJL2Dc0rO5dXsa+zzRGQ9Db1d9Y78L6roLICzy15D+Itfw7mWt0XapWCtACKTWCQEPPHEE3p/Ce/evYubNwvvFUqShAYNGsDX1xcAcP/+fcTGxkIIAUmSEBISgoCAAEs0k4jKSZIkvNSnEWavuah3meCS0wC9QsLh5OZlsXa9NrAJ1w0gMoHsawfs2rUL48aNg1qtxptvvonJkyejdu3axfZJTk7GihUr8NFHH0GSJPz0008YMGCAnM2sFrh2AFnaxbg0LNp1HfGpOaW2mbsOgCHBvkosebY9ewGITCBrCLhx4wbatWsHe3t7HDt2DM2bN9e7/9WrV9G5c2eoVCpERkYiNDRUppZWDwwBZG66qgOeiUrBnPWXi+0ndwCwV0hYOLY1ewGITCTrFMFPP/0UWVlZmDdvnsEAAADNmjXD66+/jjfffBOffPIJli5dKkMriUgXXdUBAz1d0CG4ePe+qQFgSLsA3H+Yi5M3U8vVrmAfJab3DWEAICoHWUPAvn37IEmSSbMHevToAQDYv3+/pZpFREXo+rR/6a90nff/41NzEJ8ar/23qQEg0NMF/u4u+PXsXZPb6e/ujLnDWyDIR8lbAETlJGsISEhIMPkYzS/3vXv3zN0cIipB16f9IG9X5Bao9A4A1MhO+sukWwAjOwTiiz26CwrpY6+Q8O8nmyLY183kY4noH7KGAHd3dyQmJuLw4cN47LHHjDrm0KFDAIBatWpZsGVEpFkLoOSbfbQJJXgb9pgAoVYjN/2+wQDgpXTEZ7tNDwABHi6YNYizAIjMQdaKgV27doUQAvPnz8eNGzcM7n/jxg0sWLAAkiShS5cuMrSQyDYZWgvAFEG9nkHTYf8uMwC4ONrBTgJSMvN1bi+LQgHMGtgYP7zwGAMAkZnIGgJee+01KBQKPHjwAB07dsTnn3+O1NTSg4HS0tLwxRdfoFOnTkhPT4ckSZg5c6acTSWyKcasBaDL/SuHkRp1rtTzZd2j91I6wK+WE1TlyBofj2mN/q0DeP+fyIxkDQEdO3bExx9/DCEEHjx4gJkzZ8LHxwchISHo3LkzunTpgpCQEHh7e+O1117TBoSFCxeiY8eOcjaVyKaUZy0AzSDAi6vm6AwCJUkAJncLRkyS7pUF9flkbBjC6nuYfBwR6Sf7KoKvvvoqGjRogH/961+4e/cuhBCIiopCdHQ0ABQrNezv74/Fixdj2LBhcjeTqForOQPAS+lo0vFFZwEIdT4S/zgKz+C2eo+ZOagJPE28DgCM6lgXrRt4mnwcERlmlaWEhw4diieffBJbtmzB/v37cfnyZe2nfg8PD7Rs2RK9e/fGkCFD4OBg/MphRGTYhbg0fF6iyl+QjysC3F1wN7105b+SdE0DbPzkv/QeE+yrRL9WfogxYZAhULhQ0dQewSYdQ0TGs0oIAAAHBweMGDECI0aMsFYTiGzO5tO38fX+W6Wej07MgkIC7CTovV9fnkqAdkUW9Wno7Qp/d2ckpOcabKudQsKcwc04BoDIgmQdE0BE8hJCIDoxE6ejUrDr/B2dAUBDLQB3V0cElbEMb3kCQIC7Cz4e10Y7ml+SJIztXN+otr86oDFnARBZmNV6AjTUajVSU1ORnZ2NOnXqwM7OztpNIqoWdBX+MSQlMx/znm6BtzZcQXKRKXymBgB/d2eM69wA/Vr5aZcV14xBaOznBv9azkh4UHZvQJCPK/q18jO63URUPlYJASqVCitXrsTKlStx5swZPHr0CJIk4dKlS2jWrJl2v+3bt+PIkSOoVasW3nzzTWs0lahKKqvwjzEOX0sqFgDyHqbg6ob/GhUAZvQPRfPAWmjo7artxtcVRgI8nKEAoNZxfXuFhJf7hvI2AJEMZA8BiYmJGDJkCE6dOgVDCxg2aNAATz31FCRJwqBBg9C6dWt5GklUhVW08M+vkXeK/duppheajfgP/lj/Ifxa99XbA+Dn7owgH6X232WFkbtpubBTSAio6VxsMCIXAyKSl6whQKVSISIiAmfOnIFCocDTTz+NJ554Ai+//LLO/Vu0aIHHHnsMp0+fxi+//MIQQGSE6MTMchX+0ch5VPrzuW/L7nB290XNOo0NlAJ2+qfr/2Euvt5fdhhRqQWcHRX4dkp7pGblw0vpVKwHgYgsT9YQ8MMPP+DMmTNwcHDA1q1b0a9fPwAoMwQAwFNPPYVTp07h2LFjcjWTqMq6GJeGT3Zcq/B5clIT4OLpX+y5WnWb6j0m2FeJhzmP8MLySKNDSHRiFiRJQniQl+GdicjsZJ0dsGbNGkiShGnTpmkDgCFt2rQBAFy/ft2STSOq8i7EpuL1ny8YNf1On/tXDuPkoomIP7XF6GPsFRL6tfLDf9ZcNLkXIiXT9GqFRGQesoaAS5cuASj8dG8sHx8fAEBKSopF2kRUHVyIS8Oc9ZfLVZO/qH9mAahwfesXRpUDDvR0wfzRrbD17J1yjUPwUjqVp6lEZAayhoD09HQAgJeX8V1/KpUKADh1kKgMmsF3+QW6xtqXzdWp+O+UrmmAHg1bGzzPgDB/fL77RrEKhMYK9lWiYRl1CYjI8mQNAZ6ehfW/b9++bfQxN28Wrjfu7e1tkTYRVWWamQCqcnwCz8pTaR+XpxCQxveHossVAOwVkraSIBFZh6whoHnz5gCAM2fOGH3MunXrCgcOhYdbqllEVVZ5lwC2V/zzxluRAOBoryjXLYhgHyUWjm3NqYBEViZrCBgyZAiEEPjqq6+QlpZmcP+NGzdi27ZtAIDhw4dbunlEVU55lgAGoL13X5EAYKeQTLoFUderBj4a1RJLp4ZjyZT2DABElYCsIeC5555DvXr18PDhQ/Tt2xdXr17VuV9iYiLefPNNjB07FpIkoUWLFhg5cqScTSWq9IQQuBb/wKRj7Ir0ACT+caTcAQAAhrSrY/R17RUSXh3QGB2CayPIR8lbAESVhKx1ApycnLBlyxZ0794dZ8+eRcuWLdG4cWPt9vHjxyMzMxPR0dEQQkAIAS8vL2zatIl/NIiKuBCXhg9++QMPsh8Z3NfeTkL/ln7YeTGh2NgB51q+sHOsgYLcTJMDAAAEerkYt5+nC14b2ISf/IkqIdlXEQwLC8OZM2fw+OOPF36SufZPYZOLFy/i1q1bUKvVEEKgQ4cOOHXqFBo1aiR3M4kqrc2nb2PWTxeMCgB2CgnzR7XC1bsPUXLsYM3Axmgz+WMEPj7M5AAAAM3q1EJwkRLButT1qoHvn+/AAEBUSVllAaFGjRrh+PHjOHbsGLZu3YrIyEgkJiZCpVLBy8sLbdq0wVNPPYU+ffpYo3lEldaF2FS9ywGX9OqAxqhZwxHRiVk6t9cMbIyagY11btPH390ZEoAX+zTCf8pYqEhzC0BhYrggIvlIwtAqPlRlxcfHo27dugAKp2UGBgZauUVUEUIITP72lEnT8SZ1a4iH2Y+w+Uw87l85jPTYywgdNN1st9eCfZTo18oPey7dKzZLgQsBEVUNVukJICLTxSRlmTwff+XhGADFZwEItQqNI/7PLEEgKjETSw9EYf7oVqhZwxEpmXlcCIioCpG1n06hUMDe3r7MWQG6REVFaY8jsmXlnQ5Ychqg+lEeYEIHYIC7E7yUjmVuL1ALfLM/Cg29XREe5MXR/0RViOw368p794F3LcjW1XYzvcZ+ResAzBrUBHOHt0RKZr7efaMSMxGTpHvcARFVXlVmxA4/WZCta+jtanA0flEVCQCBni74eGxr9A/zR7KBAKDB1QCJqp5KHwKSk5MBAK6uXGSEbJskSXipT6NiJX/LUtEAUHRan7E9EFwNkKjqsUoIMPZTfVZWFhYvXgwACA4OtmSTiKqEsPoeWDAmTG+PQEUCgL1CwmsDmxSb1mdMDwRXAySqmiw62i4oKEjn83379oWDg4PeY/Py8pCYmAi1Wg1JkhAREWGJJhJVOWH1PbBkSnusOByNn0/8VWybWlWA6P0ry90DoKuyn6YHYraeegBcDZCoarJonQBzFQnp2LEj9u3bx1sCJmKdgOpLCIHvD0VhzcnSy3LnPkjCueUz4V6/hdEBoK6nC5Y/30Hv7+zFuDR8ve8W6wEQVSMWDQGTJ08u9u8ffvgBkiThqaeegru7e9mNkiQ4OzvD398fnTp1Qs+ePfkpoxwYAqqni3Fp+HTHddxNL7tmQH7WAzi4uBldCnjWoCboH+ZvcD8hBGKSslgPgKiakLVioEKhgCRJuHz5Mpo1aybXZW0WQ0D1czEuDa//fAGqIr+1qdHnUatuM9g5lH9g3n9Ht0J4kJcZWkhEVYmsFXjeffddAICPj4+clyWqFoQQ+N/em8UCgGYQoEfDNmg14cNyBwGO7CeyTVYJAURkupikLEQXKchTdBZAatRZ3Dm9DfU6jzD5vBzZT2S7Kn2dACIqVLRssK5pgHUfH6bd7lfLuE/2dhzZT2TTZA0BJ06cgJ2dHVxcXHDnzh2D+9+5cwfOzs6wt7fH2bNnZWghUeUhhEB0YiZOR6Ug6n4GcvMLABhXB2Du8BZ4Z2gzeLrqn4rrW9MZLevWstyLIKJKTdbbAWvXroUQAk8++STq1KljcP86deogIiICmzZtws8//4x27drJ0Eoi69M1HQ8wvhDQB7/8gTtpuQavczc9BzFJWQgyoRwxEVUfsvYEHDt2DJIkYcCAAUYfM2jQIADAkSNHLNUsokrlYlwaZq+5WO4AAMCoAKDBmv9EtkvWEBAVFQUAJk0PbNKkCQDg1q1bFmkTUWUihMDX+26VqsyXGnWu3KWADeHMACLbJWsIyM0t/HTi7Oxs9DFOToV/oLKyuEwpVX8xSVmlegAAoFbdZnBvEAbAvAGAMwOIbJusIcDT0xMA8NdffxnY8x/x8fEAoLfCIFF1UXQGQFF2js4ImzAPIQNeNFsAYM1/IpI1BGhuA2zdutXoY3799VcAQOPGjS3RJKJKpeiyvWqVqtg2O0dn1OvytHl6AHyUWDi2NWv+E9k4WUPAwIEDIYTAqlWrcPToUYP7HzlyBKtXr4YkSXjyySdlaCGRdTWoXQP2Cgn3rxzG6a+eQ97DFLOe39/dGUuntMeSKe0ZAIhI3hAwbdo01K5dGyqVCgMHDsRXX32lHSdQVG5uLr788ksMGjQIBQUF8PDwwIsvvihnU4lkUbQWQHRiJmKTsnDn0iH8se4DZCXG4tzyV/Eo+4FZrmWvkPDvJ5siyNeNtwCICIDMCwgBwP79+zFw4ECo/u7qdHV1Rbt27eDvX7iCWUJCAiIjI5GdnQ0hBOzt7bFjxw706dNHzmZWC1xAqHLTVQsgP/oEjq14x+yzALjkLxHpInsIAICDBw9iwoQJuHv3bmEjSnwq0TSpTp06WL16Nbp37y53E6sFhoDKS1MLoOhUQFPqABRlbyehoMiqQsE+SrzUpxHcXBy45C8R6SVrxUCNHj16ICoqCqtWrcL27dtx/vx5JCcnAwBq166Ntm3bIiIiAuPHj9dOESSqLnTVAihvAACAApWAnQQ82z0I4cFexd7wWQmQiPSxSk8AyYM9AZVTdGImnl92RvvvigSAooJ9lFgypT0/8ROR0biKIJHMDK0GWN4xAFGJmYhJYlEtIjIeQwCRzLyUjtrH6TEXzToIkOsAEJEprDImwFbExsZi27ZtOHToEC5duoQ7d+5ArVajdu3aaN++PUaPHo0RI0bA3p4/hupOCIGYpCycjkrB7osJ2udDB70MoVZBXfDILLMAuA4AEZnCImMCgoKCCk8uSdpFg4o+Xx4lz1XZvf3225g3bx4MfXvDw8OxceNG1KtXz+xt4JiAyqGsZYE1ND0BFZ4G6KvEkmc5JoCIjGeRj6CxsbEASk/90zxfHlXtD1tCQgKEEHB1dcXQoUPRq1cvhISEwNnZGX/++Se+/PJLnDlzBmfOnEHv3r1x7tw5KJUcyV0daD71J2fkISUzD1/suqGdCZD4xxE4u/uhZp1Q7f5cB4CIrMUiIeCZZ54x6fnqyMvLCwsWLMCLL74INze3YtvatWuHMWPGYOzYsVi/fj1u3ryJzz77DO+8846VWkvmou9Tv2YQoJ2TK9pM/rhYEDCWnUKCb01n3E3P0T7HQkBEVF6cImhFKSkpCAgIQH5+Plq2bIlLly6Z9fy8HSAvXQWANErOAqjbaThCB0036fyaN/uWdWshJimLhYCIqMI4Is2KvLy80KpVK0RGRlap8Q5Umq4CQBq6pgGGDDBuLQzfmk4Y3D4Q7YM8SxUBYiEgIqoohgAry8srnNJlZ2dn5ZZQRcQkZem9BVDeaYAzBjRGh2Avs7aViEiDdQKsKDExEX/++ScAoGnTplZuDVVE0QJAGuYoBFTbjVP+iMhyLNITcOTIEUucFk888YRFzmstH3/8MQoKCgAAI0eONPn4+Ph4vdsTEhL0bifzKflmbY4AEOyrRENvV7O2k4ioKIsMDFQoFGYfqCRJkvYNszo4deoUunTpgoKCAgQGBuL69euoUaOGSecw5XvMgYGWJYTAC8sjEZWYiZzUuzi56BkIdeFy2eUJAPYKCQvHtuaIfyKyKIvdDhBCmP2rurh//z5GjBiBgoICSJKEH374weQAQJWLJEl4qU8j2CskuHgGIOTvkf+aAGCnUEBhZGYL9lEyABCRLCxyO+DgwYNlbsvPz8dbb72FM2fOwNvbGyNHjkSHDh3g6+sLoPAN8syZM1i/fj0SExMRHh6OefPmwcHBwRJNNUuPxYoVKzBp0iSj9s3IyMCgQYO0Xfnz589Hz549y3Xd27dv692ekJCADh06lOvcZLqw+h5YMCYMX++7BXQcAlfvevBo2BqN/Gpiet8Q7QyCogMIg32UeKlPI7i5OHDKHxHJT8hIrVaL/v37C4VCIaZOnSoyMzPL3DcrK0s899xzQpIkMWDAAIu1CUCFv1asWGHUtXJyckSPHj20x82aNctir0sIIW7fvq291u3bty16LVuXkJCgfaxWq0XU/QxxOipZRN3PEGq12qhtRERyk7VY0LJly/D888+jT58+2LNnj1HH9OvXD/v378c333yD559/3uxtunbtWoXP4e/vj1q1aundp6CgAMOGDcO2bdsAAFOnTsV3331X4Wvrw2JB8ti4cSMmTJiAH3/8EcOHD7d2c4iIjCZrCOjSpQtOnjyJzZs3Y/DgwUYds3XrVgwZMgSPP/44jh8/buEWWoZarca4ceOwdu1aAMCoUaPw888/Q2GGmvH6MARY3saNGzF69GioVCrY29vjxIkTCA8Pt3aziIiMImudAM2nblNWzNO8iZnjE7u1TJs2TRsAIiIi8OOPP1o8AJDlFQ0AADBhwgS0a9fOyq0iIjKerO9Eubm5AAwPaCtKs6+msl5V89prr2HZsmUAgF69emHDhg2wt2ehxqquZACYPHkyli1bxnBHRFWKrH+xGjVqBABYsmSJ0cdo9g0ODrZImyxp7ty5WLRoEQCgU6dO2LJlC5ycWAGuqmMAIKLqQtaPpCNHjsSlS5ewZ88evPTSS/jss8/g7Oysc9+8vDzMnDkTu3fvhiRJGD16tJxNrbDFixfjvffeAwDUqVMHCxcuRExMjN5jGjdubLGpkGQeDABEVJ3IOjAwNzcXbdu2xbVr1yBJEnx9fTFy5EiEh4fDx8cHkiRp6wRs2LAB9+7dgxACTZo0wfnz56vUp+ju3bvj8OHDJh0TExODBg0amK0NHBhYfkIIxCRlITkjD7XdCufub968GaNGjWIAIKJqQ9aeAGdnZxw8eBCDBg3CuXPncO/ePSxevFjnvpps0qZNG2zfvr1KBQCq2i7Gpeks6tPaCXBwcIBKpWIAIKJqQdaeAA2VSoUlS5bgm2++wdWrV3Xu07RpU7z44ot48cUXucxuObEnwHQX49Iwe81FFKhL/1rYKyQMrpOCyyd/w1dffcUAQERVnlVCQFH37t3D5cuXkZqaCgDw8PBAy5Yt4e/vb81mVQsMAaYRRRYBKkuwjxJLprRnWV8iqhasPlfNz88Pfn5+1m4GEWKSsooFgPtXDiMnOR4Nuo/TPheVmImj15PQtbE3gwARVXlWDwFElUVyxj+1KO5fOYw/1n0AoVZDCDUa9pig3fb+5j+0C/+E1fewRlOJiMzCaiFArVbj4MGDOHnyJO7du4fs7GzMmzev2G2A/Px8FBQUwM7OjgMDyeJquxX+HysaAAAgJzUBQohin/yjEjMxe81FLBgTxiBARFWWVUY2bd++HY0aNULfvn3x7rvv4ptvvsEPP/yAtLS0YvstW7YMbm5u8PHxQVZWljWaSjakobcrpNunigUA/7b90XToLJ1d/wXqwqWBrTyshoio3GQPAd999x0GDx6M2NhYCCHg5eVV5h/RqVOnolatWsjMzMQvv/wic0vJ1mzatAkHv3urdADQMwsgKjETMUkMqERUNckaAm7evInp06cDAHr27ImrV68iMTGxzP0dHR0xfPhwCCGwd+9euZpJNkQIgejETPz3qxXFKgGGdo4wGAA0UjKr5roWRESyhoBFixahoKAAzZs3x86dO9GkSRODx3Tt2hUAcP78eUs3j2zMxbg0vLA8EoNf+QRvzpiqDQCDnx6Lq4d/wbsjWhp1Hi8lx6sQUdUkawg4cOAAJEnCK6+8AkdHR6OO0Sw6ZMrKg0SGaIoC3biTgps7vyl2CyCn9RRciX+Iro29Eeyj1HueYF8lGnq7ytFkIiKzkzUExMfHAwDCwsKMPsbVtfAPbHZ2tkXaRLZHiMIBfQVqATsHJ7SZvBCOSk/tGAAVJCzadR2no1IwtH0d2JVRDsBeIWF6nxDWCyCiKkvWKYKaP5amvKGnpKQAAGrVqmWRNpHtKVkUyNW7HsJf+gZObl7aMQDxqTl4c/1lAECAhzMgJNxNz9EeE+yjxPS+IWhVz13WthMRmZOsIaBOnTq4efMmoqOjtff6DTl27BgAICgoyJJNo2qi5Op/DWrXQGxydrHVAA8ePga1yh6KImtSONfyLvOcd9NyYa+QMHNQY9R2c4KXsvA87AEgoqpO1hDQvXt33LhxAz/88AOeeeYZg/s/ePAAS5YsgSRJ6NmzpwwtpKpM1+p/jvYK5Beotf8uiD2Jo9+/A+/m3dBsxBvFgoA+BWqBX8/c4boBRFStyDomYNq0aZAkCYcPH8bKlSv17puSkoIhQ4bg3r17sLe3xwsvvCBPI6lK0gz0K7n4T9EAcP/KYRxZ/jbUKhXuXzqAhHO7TboGawIQUXUjawho06YNZsyYASEEpkyZglGjRmH9+vXa7SdOnMDPP/+M6dOno1GjRjhy5AgkScLbb7+N+vXry9lUqkKKDvQrS8lSwP5t+6Nu+ECTr8WaAERUnci+lLAQAi+//DK++eYbvd2qmma98sor+Oyzz+RqXrViK0sJRydm4vllZ8rcrisANB06Cz61nOFgp8Dd9Fyjr7V0ajiCDEwbJCKqKmQvGyxJEv73v/9hz5496N69OyRJghCi2BcAPP7449ixYwcDABlUdPW/ksoKAJJCgaSMfNwzIQCwJgARVTdWW0WwT58+6NOnDzIyMnD+/HkkJiZCpVLBy8sLrVu3Ru3ata3VNKpiait1F57SFwA01DqPLI01AYioOpI1BDz77LMAgAEDBuDpp58GALi5ueGJJ56QsxlUjVyMS8P/9t4s9XzKjdMGA4CxWBOAiKorWUPADz/8AAAYNWqUnJelakozI0DXgMCagU2g9A1CRsKtCgWAGf1D8GSbOuwBIKJqSdYQ4O3tjaSkJPj6+sp5WaqGDM0IcKhRE22e/QTxp7agQbdxpQKAnUKCSs9sAo3mge4MAERUbck6MLBZs2YAgLi4ODkvS9VQydK/ALRd/xoONWriywXvo27t4oP5gn2UWDgmjIsDEZHNk7UnYPz48Th06BB++OEHDB48WM5LUzVTckbA/SuHcfvEJrSe+BHsnf95c1/7+23MHNQENV0ckJKZV6zk70t9GpV5O4EDAYnIFsjaEzB58mT06tULW7Zswdy5cyFziQKqRmq7OWkfa2YBPIi7gvMrXkdB3j8L/SSk5+I/ay4iI+cRwoO8EOSj1L6xh9X3wAIdPQLBPkosHNuaAwGJqNqTtVjQkSNHkJOTg9mzZ+Py5csIDQ3FqFGj0KpVK3h4eMDOQB13ziIwTXUuFiSEwAvLI3HiwA6jZgEE+yjLrPuvWXSoZE8BEVF1J2sIUCgU5f7jKkkSCgoKzNyi6q06hwAA+OSblXj95SlGTwNktT8iouJkLxbEWwBUXkWXCT5zaBf+86+pJtUBSMnMYwggIipC1hBw8OBBOS9H1UjRZYKNqQSoi5fSSe92IiJbI2sI6Natm5yXo2qiaFGg8gYATvcjIipN9gWEiExRsihQ4pXDJgcATvcjItJNlp6AHTt2YPfu3YiLi4NKpUJAQAC6d++OkSNHwsHBQY4mUBVS9N5/7iNVsaJAzZ+eA6EqgL2z0qgAEOjpgtcGNuF0PyIiHSw6O+D+/fsYMmQITp8+rXN7gwYN8Ouvv6Jly5aWaoJNq4qzA4re+y+LWlUASVIYFQC+f74DFOVYM4CIyBZY7K+jSqXCU089hVOnTkEIofMrJiYG/fr1Q3JysqWaQVWI5t5/0QCQdPU4slPuFNtPYWdv1C2A1wY2YQAgItLDYn8h169fjzNnzkCSJDRq1AjLly/H5cuXce3aNWzYsAEdO3YEUNhb8Omnn1qqGVRF6FoQ6P6Vw7i85l2cW/5aqSCgDyv+EREZx6IhACjs8j99+jQmT56M5s2bIzQ0FMOHD8fRo0fRrVs3CCGwYcMGSzWDqoiSCwIVnQWQ9yAJdyN3GHWeQE8XfPNsOwYAIiIjWCwEnD9/HpIkYebMmXB3dy+13c7ODu+99x4AICYmBhkZGZZqClUBRRcE0jUNMLjPVKPOE5+ag9jkbIu0kYiourFYCEhKSgIAtG/fvsx9im7juADbplkQqLx1AIpKycwzvBMREVkuBOTkFK7kplSWXaa1Ro0a2se5ubmWagpVAQ29XSHdPlXhAACwMiARkbEqzdBprilg2zZt2oSD371V4QDAyoBERMaTfQEhopKuXLmC0aNHQ6VSAQBCO0cgsP8MbQBwtFcgv0Bt8DysDEhEZBqLh4Cvv/4aPj4+ZtnvnXfeMVezqBJp3rw5/vOf/2DevHmYPHkyvvvuO8Sl5CAlMw9eSic0qF0DscnZ2n8/zHmEb0oUFAr2UWJ63xDOCiAiMoHFKgYqFAqzfyLTfFIk41SlioFCCPzyyy8YMmSIUQV+NKWFNcGgobcrewCIiExk0RBgTpIkMQSYqDKHgLS0NHh4eFi7GURENs1itwMOHjxoqVNTFbdx40ZMnToVW7Zs4fLSRERWZNEFhMi6KmNPwMaNG7WDAGvUqIEzZ86gWbNm1m4WEZFNqjRTBKn6KxoAAGDUqFFo0qSJlVtFRGS7GAJIFiUDwIjR4/Hdd99xlT8iIiviX2CyuJIBwL9tf6Q2n4SXVpzDxbg0K7eOiMh2MQSQRekKAJpKgFGJmZi95iKDABGRlTAEkMXoCwAaBWqBr/fdYtloIiIrYAggi8nPz9e+uetbCyAqMRMxSVlyN4+IyOZx7QCymLFjxyLqfgaW/LwVTQa/pncxoJTMPAT5lL3iJBERmR9DAFnUuHHjcDgn1OB+XP6XiEh+vB1AZrNx40YsX7682HMNvV0RbOATPpf/JSKyDvYEkFmUHAQ4ZcoUAIVrPrzUpxFmr7mIAnXpwX9c/peIyHrYE0AVVjIAnDx5stj2sPoeWDAmrFSPQLCPEgvHtubyv0REVsK1A6oxOdYOKBkAJk+ejGXLlumsBMjlf4mIKhfeDqByMyUAAIW3BoJ8lJwFQERUSfB2AJWLqQGAiIgqH/7FJpMxABARVQ/8q00mefjwIaZNm8YAQERUDfAvN5mkZs2a2LFjB2rWrMkAQERUxXFgIJmsY8eOiIyMRHBwMAMAEVEVxr/gZNCFCxdKrfIXEhLCAEBEVMXxrzjptWHDBrRv3x5jn30RUfcztGFACIHoxEycjkpBdGImlwImIqqCeDuAyvTJNyvx+stTIdQqrF35La7k+KBT7wj0a+WHPZfuISoxU7tvsI8SL/VphLD6HlZsMRERmYI9AaRTYQCYAqEunAXg37Y/fFp0Q1RiJr7ef6tYAACAqMRMzF5zERfj0qzRXCIiKgeGACplw4YNf/cAqAEUBoCmQ2dBMjAGoEAt8PW+W7w1QERURTAEWMGuXbsgSZL2a+7cudZuktbGjRsxZsyYYj0AxgQAjajETMQkZVmyiUREZCYMATLLysrCiy++aO1m6FSyEqCpAUAjJTPPEs0jIiIzYwiQ2dtvv424uDj4+PhYuynFbNu2zSwBAAC8lE7mbh4REVkAQ4CMzp49iy+//BJOTk6YN2+etZtTTFhYGOrXrw8AmDRpEp6cNrdcASDYV4mG3q7mbh4REVkAQ4BMVCoVnnvuOahUKsyZMweNGjWydpOKqVevHg4ePIg5c+Zg+fLlmN4vFPYKyaRz2CskTO8TAkky7TgiIrIOhgCZLFq0COfPn0doaChmz55t7eYAQKlR/PXq1cO8efOgUCgQVt8DC8aEIdhHWWwfTT0AXc8vHNsareq5W7rZRERkJiwWJIPY2Fi8++67AIBvvvkGTk7Wv2e+ceNGrFq1CuvXr4ezs7POfcLqe2DJlPaIScpCSmYevJROaOjtCkmSMLR9oM7niYio6mAIkMGLL76I7OxsjBs3Dj179rR2c4rNAhg2bBh++eWXMoOJJEkI8lEiqMQn/7KeJyKiqoMhwMJ+/vln7N69G+7u7vjss8/Meu74+Hi92xMSEko9V3IaoJ+fHxwcHMzaLiIiqhoYAiwoNTUVr776KgDgv//9r9mnBdatW9ek/UsGgMmTJ2PZsmVcDZCIyEbxr78FzZo1C4mJiXjsscfw/PPPW7UtO3bsYAAgIqJibL4nwByD2VasWIFJkyYVe+7QoUNYsWIF7OzssGTJEou82d6+fVvv9oSEBHTo0AEA8NJLL0H991oADABERAQwBFhEXl4epk2bBgD4v//7P7Ru3doi1wkMDDR6XwYAIiIqyeZDwJ9//lnhc/j7+xf79+bNm3Hjxg04ODigWbNmWLt2baljrl69qn185coV7T6PPfYYGjZsWOE26cIAQERERUmC676a3cqVKzF58uRyHavr1kJ5xcfHawcPjhw5EmvWrGEAICIiLb4j2IiPP/6YAYCIiIrhu4IFTJo0CUIIvV8HDx7U7v/uu+9qnzdXL0BJDABERFSSzY8JqM4KCgq0j3UVDiIioqrJz88P9vYVfwtnCKjGkpKStI81UwWJiKjqu337tkkzxMrCPmIiIiIbxZ6Aaqxly5Y4ffo0AMDb29ssXUfVRdFCSqdPny41zZMqP/4Mqz7+DMvPz8/PLOfhu4KVdO/eHZaenens7Izw8HCLXqM68Pf3N0u3GlkPf4ZVH3+G1sHbAURERDaKIYCIiMhGMQQQERHZKIYAIiIiG8UQQEREZKMYAoiIiGwUQwAREZGN4lLCRERENoo9AURERDaKIYCIiMhGMQQQERHZKIYAIiIiG8UQQEREZKMYAoiIiGwUQwAREZGNYgggIiKyUQwBRERENoohgKq92NhYLF68GMOHD0dISAhq1KgBZ2dnBAYGYsiQIVi7di0KCgoqfJ1Dhw5BkiSjvubOnVvxF1ZNxMXFYebMmWjSpAlcXV3h6emJ8PBwfPzxx8jOzjbbdXbt2oWhQ4ciMDAQTk5OCAwMxNChQ7Fr1y6zXcOWREZG4v3330ffvn2131OlUonQ0FBMnjwZx44dM8t15s6da/Tv1aFDh8xyTZsiiKqxt956S0iSJADo/QoPDxdxcXEVutbBgwcNXkfz9e6775rnBVZxW7duFTVr1izz+xQaGipu3rxZoWuoVCoxZcoUvT+PqVOnCpVKZaZXVf117drVqP/nEydOFHl5eRW61rvvvmv079XBgwfN8wJtiL25QwVRZZKQkAAhBFxdXTF06FD06tULISEhcHZ2xp9//okvv/wSZ86cwZkzZ9C7d2+cO3cOSqWywtf9/vvvER4eXuZ2Hx+fCl+jqjt//jxGjRqFnJwcKJVKvPHGG+jRowdycnKwdu1afPfdd7hx4wYGDRqEyMhIuLm5les6b775JpYvXw4AaNOmDV5//XUEBwcjKioKCxcuxPnz57Fs2TJ4e3vjo48+MudLrLbu3r0LAAgICMDTTz+Nrl27ol69elCpVDh58iQ+/fRT3LlzB6tWrcKjR4/w888/m+W6ly9f1ru9YcOGZrmOTbF2CiGypNdff10sWLBAPHz4UOf2goICMXLkSO0niffee6/c1yraE8BPJIZpPk3a29uLEydOlNq+cOHCCvecXL9+Xdjb2wsAon379iI7O7vY9qysLNG+fXttOyra62ArBg0aJNatWycKCgp0bk9KShKhoaHan9/hw4fLfa2iPQFkfvyuks1LTk4Wjo6OAoBo2bJluc/DEGC8U6dOab9X06ZN07mPSqUSTZs2FQCEu7u7yM/PN/k6L774ovY6J0+e1LnPyZMntfu89NJLJl+DdNu2bZv2+/qvf/2r3OdhCLAsDgwkm+fl5YVWrVoBAKKioqzcGtvw66+/ah9PnjxZ5z4KhQITJ04EAKSnp+PgwYMmXUMIgS1btgAAmjRpgo4dO+rcr2PHjmjcuDEAYMuWLRBcXd0sevTooX3M36vKiyGACEBeXh4AwM7OzsotsQ2akeOurq5o165dmft169ZN+/j48eMmXSMmJkZ777roefRd586dO4iNjTXpOqSb5ncK4O9VZcYQQDYvMTERf/75JwCgadOmZjnnm2++ifr168PJyQkeHh5o06YNXn31Vdy4ccMs56/qNN/vRo0awd6+7PHJTZo0KXWMsa5evarzPOa+Dul2+PBh7WNz/V717dsXPj4+cHR0hI+PD7p374758+cjLS3NLOe3RQwBZPM+/vhjbZ2AkSNHmuWcJ06cwF9//YX8/Hykp6fjwoUL+Pzzz9G0aVPMnTvXprucc3NzkZycDAAIDAzUu6+HhwdcXV0BALdv3zbpOvHx8drHhq5Tt25d7WNTr0OlqdVqzJ8/X/tvc/1e7du3D0lJSXj06BGSkpJw+PBhvPHGGwgKCtLe+iHTcIog2bRTp07h888/B1D4RvHiiy9W6Hz+/v4YNmwYunTpgqCgINjb2+Ovv/7C9u3btdOl3nvvPeTn59vsdLSMjAztY2OmY7q6uiIrKwuZmZkWu44maAAw+TpU2qJFi3D69GkAwLBhw/Te8jFGy5YtMWTIEHTo0AEBAQF49OgRrl+/jp9++gl79+5Feno6hg8fjm3btmHAgAHmeAm2w8oDE4ms5t69eyIwMFAAEJIkid9++61C58vMzNQ7gv3UqVOiVq1a2utduHChQterqv766y/taO8JEyYY3L9u3boCgAgODjbpOu+//772OoZ+tr/99pt23w8++MCk61Bxhw4d0k7L9PHxEffv36/Q+dLS0vRuX7JkifZnFxAQIHJycip0PVvD2wFUKRhbFlTf18qVK42+XkZGBgYNGqTtMp4/fz569uxZodfg6uoKBweHMrd36NABX331FYDCkeuax7bG2dlZ+zg/P9/g/poBZi4uLha7TtFBbKZeh/7xxx9/YOjQoSgoKICzszM2bNhQ4cJY7u7uerdPmzYNU6ZMAVBYxGjTpk0Vup6tYQggm5Obm4vBgwfj7NmzAIBZs2bh9ddfl+Xao0ePRs2aNQEUHzhlS4pW/jOm6z0rKwuAcbcOynsdzTXKcx0qFBMTg759+yItLQ12dnZYu3YtnnjiCVmuPW3aNO1jW/29Ki+OCaBKwRwjsv39/Q3uU1BQgJEjR2rnnE+dOhUff/xxha9tLHt7e4SGhiIyMhJ37tyR7bqVibOzM7y8vJCSklJs8J4uaWlp2jfoooP3jFF0MKCh6xQdDGjqdajwE3jv3r1x9+5dSJKE77//HoMHD5bt+s2aNdM+ttXfq/JiCKBKwdAULnNQq9WYMGECtm3bBgAYNWoUvv32W4tftyRJkmS/ZmXTrFkzHD16FLdu3UJBQUGZ0wSvXbumfWzqNLOibwxFz2Pu69i65ORk9OnTB9HR0QCAxYsXa4s8yYW/U+XH2wFkM6ZNm4a1a9cCACIiIvDjjz9CoZD3V6CgoEBbKyAgIEDWa1cmXbp0AVDYDa+5LaNL0a7dzp07m3SNhg0bar/HhrqIjxw5AgCoU6cOGjRoYNJ1bNmDBw/Qr18/bU2G+fPnY/r06bK3o2hNCFv+vSoPhgCyCa+99hqWLVsGAOjVqxc2bNigt0iNpaxbtw4PHjwAYLiKXXU2ZMgQ7eMVK1bo3EetVmPVqlUACgeHFS1DawxJkrRd0teuXcPvv/+uc7/ff/9d2xMwePBgfqo0UnZ2NgYNGoRz584BKCyQNXv2bKu0pWiPni3/XpWLtacnEFla0QVIOnXqJDIzM00+R0xMjPYc3bp1K7U9NTXV4KJBp06dEu7u7topgpGRkSa3ozqp6CqCRRdseuaZZ3Re4/r168LOzq7MVQSzs7OLrSJ448YNc7y0ai8vL0/07dtX+/2fMWNGuc6zYsUKvT/jS5cuGVzZ8dtvv9Wew8/Pr1y/37aMYwKoWlu8eDHee+89AIVdvQsXLkRMTIzeYxo3bqx3qp8uDx48QI8ePdCqVSsMGTIE7dq1g7+/P+zs7LTFglavXq2dqjZr1qwKF1Cp6r744gt07twZOTk56Nu3L+bMmYMePXogJycHa9euxdKlSwEAoaGhmDlzZrmuERoain//+9+YP38+IiMj0blzZ8yePRvBwcGIiorCggULcP78eQDAv//9b4SEhJjt9VVnY8aMwd69ewEAPXv2xJQpU3DlypUy93d0dERoaKjJ1zl79iymTp2KHj16YMCAAWjZsiW8vLxQUFCAa9euaYsFAYXrEyxdurRY4ScygrVTCJEldevWTfspwdivmJiYUucx1BNQdLu+Lzs7OzF37lyhVqst/+KrgK1bt4qaNWuW+f0KDQ0t85OgMT0BQhQuSfzss8/q/blMmTJFqFQqC73K6sfU36n69evrPI+hnoCi2/V9eXl5iV9//dWyL7qaYk8AkRkEBARgw4YNOHnyJE6fPo07d+4gOTkZubm5qFWrFho3bozu3btj6tSpHHhWREREBC5duoQvvvgCO3bsQHx8PBwdHdGoUSM8/fTTePnll1GjRo0KXUOhUGD58uUYPnw4li5dijNnziA5ORm1a9dGeHg4pk2bxlKzldTAgQOxfPlynDx5EufPn8f9+/eRkpICIQQ8PT0RFhaG/v37Y9KkSdr6G2QaSQgbXsmEiIjIhnF2ABERkY1iCCAiIrJRDAFEREQ2iiGAiIjIRjEEEBER2SiGACIiIhvFEEBERGSjGAKIiIhsFEMAERGRjWIIICIislEMAURERDaKIYCIiMhGMQQQERHZKIYAIiIiG8UQQEREFrV06VKEhITAyckJYWFh2L59u7WbRH+ThBDC2o0gIqLq6dNPP8WsWbOKPSdJEn755RcMHjzYSq0iDYYAIiKyiOzsbPj4+EClUmHhwoXo0KEDdu3ahffffx/BwcG4efOmtZto8+yt3QAiIqqerly5gqysLLzxxhv417/+BQB47LHHcP36daxduxZJSUnw9va2cittG8cEEBGRRTg5OQEAatWqVex5d3f3YtvJehgCiEiv48ePY+jQofDz84O9vT0kSYIkSUhPT7d206iSa968Ofz8/PDZZ5/h0KFDyMrKwubNm7Fy5Up06tQJNWvWtHYTbR5DANm82NhY7RtbRb6qo23btqFbt2749ddfcf/+fahUKms3qUrq27cvJEnCjBkzZLlebm4u3N3dIUkSGjRoAFOHfo0dO1b7//r8+fNl7vfRRx9BkiSEhYXp3G5vb4+lS5ciPT0dPXr0gFKpxPDhw+Hk5ISvvvrKpDaRZTAEEFGZZs6cCZVKhYCAAKxatQpnz57F5cuXcfnyZX6KM1JGRgYOHz4MAIiIiJDlms7Oznj66acBAHFxcTh69KjRx2ZkZODXX38FALRo0QJt2rQpc99t27YB0P+6IiIi0L1792LPjR49Wu95ST4cGEg2r06dOrh8+XKZ21u2bAkAaN++PVasWCFXs6zur7/+0o7enjNnDiZMmGDlFlVNe/bsQX5+PmrWrIlu3brJdt2JEydi2bJlAIDVq1fjiSeeMOq4TZs2IScnR3uOsiQmJuL06dMA9IeAxMREHDhwAADg6uqKrKwsbNq0CYsXL4aDg4NRbSLLYQggm+fg4IAWLVoY3M/V1dWo/aqLO3fuaB+HhoZasSVVm+bTcr9+/WR90+vSpQsaNmyImJgYbNiwAV999ZVRA/FWr14NALCzs8O4cePK3G/Hjh1Qq9Xw9fVFhw4dytxv7dq1KCgoAAAsWrQIzz//PJKTk7F7927ZekaobLwdQEQ65eXlaR/zE1v5qNVq7Ny5EwDw5JNPynptSZK0vTcPHjzQhhF94uPjcejQIQBAr169EBAQUOa+mvMNHDhQ75gYTajo2LEjpkyZoj2n5nmyLoYAogro3r07JEnS3vO8efMmXn75ZYSEhKBGjRqQJAmxsbEAgJUrV2oHW2me06XoQMWVK1fqvf6vv/6Kp59+GvXq1YOzszPc3d3Rvn17vPfee0hLSyvXa5o0aRIkSUKPHj20z/Xo0aPYIEhd7bp8+TKef/557Wt3c3ND8+bN8eqrr5r8ejdv3oyBAwciICAA9vb2pe4pa46bPXs22rVrBy8vLzg4OKB27dro2rUr5s6di+jo6DKv+eDBA/z3v/9F586d4e3tDUdHR/j7+yMiIgIbN240eSBdWX7//XckJydDoVBg4MCB5T5Pbm4uvvrqK/Tq1Qt+fn5wdHSEj48PevfujeXLl2s/aZdUtDvfmDfdn376CWq1utSxJeXl5WHfvn0A9N8K+PPPPxEZGQmgcLChQqHAqFGjABSGCM4wqQQEEekFQAAQ3bp1K7WtW7du2m2//vqrcHV11e6v+YqJiRFCCLFixYpSz+kSExOj3W/FihU690lNTRU9e/Ysda2iXz4+PuLkyZMmv95nnnlG73l1teujjz4SCoWizP2dnJzEDz/8YPD1fv/992LChAmlji/5vf/444+Fg4OD3jbq+nkJIcT+/fuFl5eX3mMHDhwoMjIyTP7elfSf//xHABCdO3cu9zkuXLgg6tevr7e94eHh4t69ezqP79SpkwAgHBwcRHJyst5rNW/eXAAQbm5uIisrq8z9du/erf256vs+vfHGGwKAsLOzE/fv3xdCCHHmzBltu5cuXWrEd4AsiSGAyABjQkDDhg2FUqkU3t7eYv78+eL48ePi999/F4sXLxZJSUlCCPOFgNzcXNG2bVvtH9cJEyaINWvWiN9//10cPXpUzJs3T/sm5+HhIWJjY016vfHx8eLy5cvi+++/L/bmfPnyZe1XWlqadv///e9/2v28vb3FJ598Ik6ePCmOHTsm5s6dqw1GkiSJHTt26H29rVq1EgBE165dxc8//ywiIyPF/v37xbJly7T7v//++9r93d3dxZw5c8S+ffvEuXPnxIEDB8Qnn3wiOnXqJLp3717qWseOHdOGB19fX/Hhhx+Kbdu2ibNnz4pt27aJ8ePHa889bNgwk75vumjeVOfPn1+u42/evClq1aolAIiaNWuKN954Q/zyyy8iMjJS7NmzR0yfPl3Y29sLAOKxxx4T+fn5pc6xZMkS7Wv66quvyrzWuXPntPtNmjRJb7umT58uAIj+/fuXuY9arRb16tUTAES/fv2KbQsNDdX+nMm6GAKIDDAmBAAQAQEBIi4urszzmCsEzJkzR/sGGBkZqfMcsbGxwt/fXwAQY8eONfQSdTp48KC2HQcPHtS5T2JioqhRo4b29f/111+l9jl37pw2CNSpU6fUG1XR1wtATJw4UajVap3XO3funLbHITQ0VNy+fbvM9pdsS35+vmjQoIH2zausT7pLly7VtmXv3r1lnt+Q6Oho7XmuXLlSrnNoPsW3adNGGyZL2rVrl/Z7ouuTdVpamnBychIARMeOHcu81quvvqpt74EDB/S2S9MzoS9UFP3/U7IXaO7cudpgqO93gSyPIYDIAGNDwKpVq/SexxwhICMjQ/vJcPHixXqv9/XXX2u7gTMzM/Xuq4sxIWDBggXafdauXVvmuT788EPtfuvXry+2rejrdXd3Fw8fPizzPGPGjNG+eZw7d86k17Nq1SoBQDg7O4vExES9+3bo0KFCAUoIIb788kttL1F5HDlyRPt9uXTpkt59R44cKQCITp066dw+YsQI7blu3rxZantBQYHw8/MTAES9evXKDGFCCHHp0iXtufSF3meffVYAEC4uLqV+pjdu3NCe44MPPtD72siyODCQyAwcHR21xVks6fDhw3jw4AEAYMSIEXr31cwLf/ToEc6ePWuR9uzfvx9AYS34YcOGlbnf1KlTSx2jS0REBNzc3HRuU6vV2LVrF4DCAZmmFpvZunUrAKBbt24GF63RfO9Onjxp0jWKMqaQjj6a9jZu3Fhbq6IsmvaeOXNG5yDBZ555RvtY1wDBffv24d69ewCA8ePH6x3tr3ldrVq1Qr169XTuk5ubi40bNwLQ/TMNCQlBeHh4me0h+TAEEJlBSEgInJ2dLX4dzUhrAPD399dbxrhoTQPNH3hzu3LlCgCgbdu2eqcR+vr6okGDBsWO0aVVq1ZlbouJidGOJu/atavJbdV87/bs2WOwBPQnn3wCoPzfN3NUCdS09/r16wbb+/LLLwMoDHypqamlztW/f3/4+PgAKJwBUFLRN2JDRaGMCTdbtmzBw4cPARTOCtBF8/yNGze0RYdIfgwBRGbg4eEhy3USExPLdVx2draZW1JI84ajeYPRx8/Pr9gxuuj7PiYnJ2sf+/v7G9tErfJ87zSV80ylqRLo5uZW7iqB5vxZ29vbY8yYMQCAqKgonDhxQrstMzNTWyY4PDwcTZo00dsmzRu2vroHmlDh4eGBAQMG6Nxn9OjRsLOzK7Y/yY8VA4nMQPPHzNKKLuBz7tw5o4v4BAYGWqpJAGC2BZQs+X3UfO8GDBiAhQsXWuw6ALB9+3YAFasSqGlvWFgYfvzxR6OPq1Onjs7nJ06ciC+++AJA4Ztup06dABSWCdYEB321AQBg586dUKvV8PHxKbNKYGJiIvbs2QMASEtLM6pK4dq1a/HZZ5+xKJUVMAQQyUSh+KfjTVOQRZesrKwyt3l5eWkfe3t7W/zN3RBPT08kJCTg/v37BvfVdK17enqW61q1a9fWPk5ISDD5eC8vL9y9exf5+fkWLf9ctEpgRcrian7WmZmZZmlv27Zt0bx5c/zxxx9Yv349vvjiCzg6Omo/hTs4OGh7C8qiuRUwaNCgYv+fi1qzZk2ZxYvKkpycjF27duGpp54y6TiqOIYAIpkUHRylr5rfjRs3ytxWdDDc8ePHtdXXrKVFixZISEjAuXPnUFBQAHt73X9SEhMTERcXpz2mPBo2bAh3d3ekp6fjyJEjJh/fpk0b3L17F5GRkcjPz4ejo2O52mHI77//jqSkpApXCWzTpg1OnDiB6Oho3Lt3T3s7pSImTpyI2bNnIzU1FTt37kR4eDgOHjwIoLD8b9GQWVJeXh727t0LQH+40YSK+vXrY/78+Qbb9MILL+DBgwdYvXo1Q4AVcEwAkUwaNmyofVx0gF9Ja9asKXNb7969UaNGDQDAl19+abbytuXVu3dvAEB6ejo2b95c5n7Lly/XtlVzjKkUCgUGDRoEoHCWhL517nXRvME8ePDAoqtBaj4tP/7448V6L0ylaa8QQtuNX1Hjx4/XfoJfvXq10WWCAeDQoUPIzMyEk5MT+vTpo3OfP//8UzsT5emnn8bo0aMNfmnGFrCMsJVYeYoiUaUHI+oElFWitqhHjx4JT09PAUC0aNFC5Obmltpn3bp1xQrn6CoW9Prrr2u3z5gxQ6hUqjKvee/ePfHdd98ZbJsuphYLCgwMFPHx8aX2uXDhglAqlUYVCyqrTLLG+fPnjS4WVHJbbm6uqFu3rgAglEqlOHz4sN5rHT16VBw6dEjvPrpoqgT+97//NfnYkjT1Cuzs7MS6dev07nvp0iWxdetWg+fs06ePACAcHR1Fo0aNBADh6ekp8vLy9B6nqRJYsvpfUZoywQDE77//brAtQgjxyy+/aI9hGWH5MQQQGWCuECBE8T+SnTp1Er/++qs4d+6c2LVrl3j22WeFQqHQVokr600xNzdXPPbYY9p9wsLCxFdffSWOHTsmzp8/Lw4cOCAWL14sBg8eLBwdHUW7du3K9bqNCQFCFC8b7OvrKxYtWiROnToljh8/Lt577z1tADCmbLChECCEEB988EGx4kJvvvmm2L9/vzh//rw4ePCgWLRokejatavOssEnT57UVs+zs7MT48aNExs2bBCRkZHi9OnTYsuWLeKdd94RLVu2NKogk77XUt4qgUXdunVLGxwBiIiICPHjjz+KU6dOicjISLFz504xb9480bFjRwFAzJw50+A5V69eXSxoAhAvvviiweMMVQksWia4bt26egsOFZWdna39P8IywvJjCCAywJwhICsrS/sHW9dX9+7dxZUrVwy+KT58+FAMGzaszPMU/erRo0e5XrexIUAIIebNm2eWBYSMCQGa62lq5pf1VdbP5OTJk9oeAUNfZbW5LBWtEqjL9evXRYsWLYxq73vvvWfwfFlZWdo3Xc2XoYWmjKkSWPT/y4wZM0x6jZqKhywjLD+OCSCSUY0aNXDgwAHMmzcPLVu2hIuLC2rWrInw8HB89dVX2L9/P1xdXQ2ex83NDZs2bcLRo0cxdepUNG7cGG5ubrC3t4enpyfCw8Mxffp07Ny5U7vkqyXNmTMH58+fx3PPPYfg4GC4uLjA1dUVTZs2xYwZM3Dt2jWD95xNvd7Vq1fxyiuvoEWLFqhZsybs7e3h7e2Nbt264cMPPyxz7nnHjh1x8+ZNLFmyBIMGDUJAQAAcHR3h7OyMunXrom/fvpg3b1652lzRKoG6hIaG4sKFC/j5558xfPhw1KtXDy4uLtrlj7t374633noLZ8+exTvvvGPwfDVq1ChWbTIkJAQdO3bUe4wxVQKLfr+HDx9uzEsrtb8QgjUDZCYJYeWRRURE1UBGRgZq166N/Px87N27t8zBc1XR448/jt9//x1z5szBvHnzrN0cMiP2BBARmcHevXsrXCWwMkpKStJWCTRnDwdVDqwTQERkBm5ubnj33XfRsGFDi9UgsIa0tDS8/fbbsLOzK7NKIFVdvB1ARERko3g7gIiIyEYxBBAREdkohgAiIiIbxRBARERkoxgCiIiIbBRDABERkY1iCCAiIrJRDAFEREQ2iiGAiIjIRjEEEBER2SiGACIiIhvFEEBERGSjGAKIiIhsFEMAERGRjWIIICIislEMAURERDaKIYCIiMhGMQQQERHZKIYAIiIiG8UQQEREZKMYAoiIiGzU/wM7ke/9nQcHswAAAABJRU5ErkJggg==",
"text/plain": [
""
]
},
"metadata": {
"image/png": {
"height": 256,
"width": 256
}
},
"output_type": "display_data"
}
],
"source": [
"from ase import Atoms\n",
"from graph_pes.graph_pes_model import GraphPESModel\n",
"\n",
"\n",
"def check_forces(model: GraphPESModel, structure: Atoms):\n",
" # make an ASE calculator\n",
" calculator = model.ase_calculator()\n",
"\n",
" force_predictions = calculator.get_forces(structure)\n",
" force_labels = structure.arrays[\"forces\"]\n",
"\n",
" plt.figure(figsize=(2.5, 2.5))\n",
" plt.scatter(force_labels.flatten(), force_predictions.flatten(), s=6)\n",
" plt.axline((0, 0), slope=1, color=\"black\", ls=\"--\", lw=1)\n",
" plt.gca().set_aspect(\"equal\")\n",
" plt.xlabel(\"True force / eV/Å\")\n",
" plt.ylabel(\"Predicted force / eV/Å\");\n",
"\n",
"\n",
"check_forces(mp0, dataset[0])"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"To demonstrate the architecture-agnostic nature of ``graph-pes``, we can swap between the `MACE-MP-0-small` and `Orb-v3-small` models very easily:"
]
},
{
"cell_type": "code",
"execution_count": 6,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"9443981"
]
},
"execution_count": 6,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"from graph_pes.interfaces import orb_model\n",
"\n",
"orbv2_xs = orb_model(\"orb-d3-xs-v2\").eval().to(device)\n",
"sum(p.numel() for p in orbv2_xs.parameters())"
]
},
{
"cell_type": "code",
"execution_count": 7,
"metadata": {},
"outputs": [
{
"data": {
"image/png": "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",
"text/plain": [
""
]
},
"metadata": {
"image/png": {
"height": 256,
"width": 259
}
},
"output_type": "display_data"
}
],
"source": [
"check_forces(orbv2_xs, dataset[0])"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"It is explicitly not the point here to compare these two models. Instead, we are just demonstrating that these two models, created by different teams and originating from different packages, can be treated identically within the confines of ``graph-pes``.\n",
"\n",
"We now know how to use these foundation models out of the box - they behave just like any other ``graph-pes`` model!\n",
"\n",
"## Fine-tuning\n",
"\n",
"Often we want to fine-tune foundation models on specific datasets to further improve their accuracy. `graph-pes` makes this easy!\n",
"\n",
"We saw above that the *force* predictions for the foundation models were already in close agreement with the DFT labels of the `SiO2-GAP-22` model. This is despite the fact that these foundation models were trained on data labelled with different functionals.\n",
"\n",
"However, different functionals tend to have different \"reference\" energies, *i.e.*, the energy of an isolated atom of element $X$ will have some non-0 energy, $\\varepsilon_X$ and for functional (a) vs (b), $\\varepsilon_X^{(a)} \\neq \\varepsilon_X^{(b)}$.\n",
"\n",
"To demonstrate this, we create an energy parity plot below:"
]
},
{
"cell_type": "code",
"execution_count": 8,
"metadata": {},
"outputs": [
{
"data": {
"image/png": "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",
"text/plain": [
""
]
},
"metadata": {
"image/png": {
"height": 308,
"width": 376
}
},
"output_type": "display_data"
}
],
"source": [
"from graph_pes.utils.analysis import parity_plot\n",
"\n",
"parity_plot(\n",
" mp0,\n",
" dataset[:20],\n",
" property=\"energy_per_atom\",\n",
" units=\"eV/atom\",\n",
" c=\"crimson\",\n",
" label=\"MP0\"\n",
")\n",
"parity_plot(\n",
" orbv2_xs,\n",
" dataset[:20],\n",
" property=\"energy_per_atom\",\n",
" units=\"eV/atom\",\n",
" c=\"royalblue\",\n",
" label=\"Orb\"\n",
")\n",
"plt.xlabel(\"Ground truth SCAN Energy (eV/atom)\")\n",
"plt.ylabel(\"Predicted Energy (eV/atom)\")\n",
"plt.legend(frameon=True);"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"`graph-pes` provides automated functionality to correct for exactly these kind of differences:"
]
},
{
"cell_type": "code",
"execution_count": 9,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"[graph-pes INFO]: \n",
"Attempting to automatically detect the offset energy for each element.\n",
"We do this by first generating predictions for each training structure \n",
"(up to `config.fitting.max_n_pre_fit` if specified). \n",
"This is a slow process! If you already know the reference energies (or the\n",
"difference in reference energies if you are fine-tuning an existing model to a\n",
"different level of theory), \n",
"we recommend setting `config.fitting.auto_fit_reference_energies` to `False`\n",
"and manually specifying a `LearnableOffset` component of your model.\n",
"\n",
"See the \"Fine-tuning\" tutorial in the docs for more information: \n",
"https://jla-gardner.github.io/graph-pes/quickstart/fine-tuning.html\n",
"\n",
"[graph-pes WARNING]: \n",
"We are attempting to guess the mean per-element\n",
"contribution for a per-structure quantity (usually\n",
"the total energy). \n",
"\n",
"However, the composition of the training set is such that \n",
"no unique solution is possible. \n",
"\n",
"This is probably because you are training on structures\n",
"all with the same composition (e.g. all structures are\n",
"of the form n H2O). Consider explicitly setting the\n",
"per-element contributions if you know them, or\n",
"including a variety of structures of different\n",
"compositions in the training set.\n",
"\n",
"[graph-pes INFO]: \n",
"Attempting to automatically detect the offset energy for each element.\n",
"We do this by first generating predictions for each training structure \n",
"(up to `config.fitting.max_n_pre_fit` if specified). \n",
"This is a slow process! If you already know the reference energies (or the\n",
"difference in reference energies if you are fine-tuning an existing model to a\n",
"different level of theory), \n",
"we recommend setting `config.fitting.auto_fit_reference_energies` to `False`\n",
"and manually specifying a `LearnableOffset` component of your model.\n",
"\n",
"See the \"Fine-tuning\" tutorial in the docs for more information: \n",
"https://jla-gardner.github.io/graph-pes/quickstart/fine-tuning.html\n",
"\n",
"[graph-pes WARNING]: \n",
"We are attempting to guess the mean per-element\n",
"contribution for a per-structure quantity (usually\n",
"the total energy). \n",
"\n",
"However, the composition of the training set is such that \n",
"no unique solution is possible. \n",
"\n",
"This is probably because you are training on structures\n",
"all with the same composition (e.g. all structures are\n",
"of the form n H2O). Consider explicitly setting the\n",
"per-element contributions if you know them, or\n",
"including a variety of structures of different\n",
"compositions in the training set.\n",
"\n"
]
},
{
"data": {
"image/png": "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",
"text/plain": [
""
]
},
"metadata": {
"image/png": {
"height": 308,
"width": 385
}
},
"output_type": "display_data"
}
],
"source": [
"from graph_pes.utils.shift_and_scale import add_auto_offset\n",
"\n",
"adjusted_mp0 = add_auto_offset(mp0, dataset[:20])\n",
"adjusted_orb = add_auto_offset(orbv2_xs, dataset[:20])\n",
"parity_plot(\n",
" adjusted_mp0,\n",
" dataset[:20],\n",
" property=\"energy_per_atom\",\n",
" units=\"eV/atom\",\n",
" c=\"crimson\",\n",
" label=\"MP0\"\n",
")\n",
"parity_plot(\n",
" adjusted_orb,\n",
" dataset[:20],\n",
" property=\"energy_per_atom\",\n",
" units=\"eV/atom\",\n",
" c=\"royalblue\",\n",
" label=\"Orb\"\n",
")\n",
"plt.xlabel(\"Ground truth SCAN Energy (eV/atom)\")\n",
"plt.ylabel(\"Predicted Energy (eV/atom)\")\n",
"plt.legend(frameon=True);\n"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Note the warning above! It we pass a set of structures with exactly the same composition (here $n\\cdot SiO_2$), its not possible to decouple the differences in the $\\varepsilon_O$ and $\\varepsilon_{Si}$. This means if you use this new model to predict energies on a new structure with a different compositions, e.g. $Si_5O_9$, you will get very different energies from what the ground truth can give you.\n",
"\n",
"With that in mind, lets get down to fine-tuning a model. As in our other [fine-tuning guide](https://jla-gardner.github.io/graph-pes/quickstart/quickstart.html#Fine-tuning) we start by selecting some structures to fine-tune on:"
]
},
{
"cell_type": "code",
"execution_count": 10,
"metadata": {},
"outputs": [],
"source": [
"train, valid = dataset[:20], dataset[20:25]\n",
"train.write(\"train.xyz\")\n",
"valid.write(\"valid.xyz\")"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Next, we define a config file to describe how what and how we want to fine-tune. Two vital features here are:\n",
"1. the `fitting/auto_fit_reference_energies=true` flag - this performs the above automated offset guessing before any fine-tuning takes place\n",
"2. using a relatively small learning rate "
]
},
{
"cell_type": "code",
"execution_count": 20,
"metadata": {},
"outputs": [],
"source": [
"accelerator = \"gpu\" if torch.cuda.is_available() else \"cpu\"\n",
"cutoff = mp0.cutoff.item()\n",
"\n",
"training_config = f\"\"\"\n",
"data:\n",
" train:\n",
" +file_dataset: \n",
" path: train.xyz\n",
" cutoff: {cutoff}\n",
" valid:\n",
" +file_dataset: \n",
" path: valid.xyz\n",
" cutoff: {cutoff}\n",
"\n",
"loss:\n",
" - +PerAtomEnergyLoss()\n",
" - +ForceRMSE()\n",
"\n",
"fitting:\n",
" trainer_kwargs:\n",
" max_epochs: 20\n",
" accelerator: {accelerator}\n",
"\n",
" optimizer:\n",
" +Optimizer:\n",
" name: Adam\n",
" lr: 0.0001\n",
"\n",
" auto_fit_reference_energies: true \n",
"\n",
"wandb: null\n",
"general:\n",
" progress: logged\n",
" run_id: mp0-fine-tune\n",
"\"\"\"\n",
"\n",
"mp0_config = \"\"\"\n",
"model:\n",
" +mace_mp:\n",
" model: small\n",
"\n",
"general:\n",
" run_id: mp0-fine-tune\n",
"\"\"\"\n",
"\n",
"with open(\"fine-tune.yaml\", \"w\") as f:\n",
" f.write(training_config)\n",
"\n",
"with open(\"mp0.yaml\", \"w\") as f:\n",
" f.write(mp0_config)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Now we can fine-tune. Note the use of config file stacking here - this lets us separate the config into different files, which can often be useful! (see below)"
]
},
{
"cell_type": "code",
"execution_count": 12,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"[graph-pes INFO]: Started `graph-pes-train` at 2025-04-11 16:52:15.879\n",
"[graph-pes INFO]: Successfully parsed config.\n",
"[graph-pes INFO]: Logging to graph-pes-results/mp0-fine-tune/rank-0.log\n",
"[graph-pes INFO]: ID for this training run: mp0-fine-tune\n",
"[graph-pes INFO]: \n",
"Output for this training run can be found at:\n",
" └─ graph-pes-results/mp0-fine-tune\n",
" ├─ rank-0.log # find a verbose log here\n",
" ├─ model.pt # the best model (according to valid/loss/total)\n",
" ├─ lammps_model.pt # the best model deployed to LAMMPS\n",
" ├─ train-config.yaml # the complete config used for this run\n",
" └─ summary.yaml # the summary of the training run\n",
"\n",
"GPU available: True (mps), used: False\n",
"TPU available: False, using: 0 TPU cores\n",
"HPU available: False, using: 0 HPUs\n",
"[graph-pes INFO]: Preparing data\n",
"[graph-pes INFO]: Setting up datasets\n",
"[graph-pes INFO]: Pre-fitting the model on 20 samples\n",
"[graph-pes INFO]: \n",
"Attempting to automatically detect the offset energy for each element.\n",
"We do this by first generating predictions for each training structure \n",
"(up to `config.fitting.max_n_pre_fit` if specified). \n",
"This is a slow process! If you already know the reference energies (or the\n",
"difference in reference energies if you are fine-tuning an existing model to a\n",
"different level of theory), \n",
"we recommend setting `config.fitting.auto_fit_reference_energies` to `False`\n",
"and manually specifying a `LearnableOffset` component of your model.\n",
"\n",
"See the \"Fine-tuning\" tutorial in the docs for more information: \n",
"https://jla-gardner.github.io/graph-pes/quickstart/fine-tuning.html\n",
"\n",
"[graph-pes WARNING]: \n",
"We are attempting to guess the mean per-element\n",
"contribution for a per-structure quantity (usually\n",
"the total energy). \n",
"\n",
"However, the composition of the training set is such that \n",
"no unique solution is possible. \n",
"\n",
"This is probably because you are training on structures\n",
"all with the same composition (e.g. all structures are\n",
"of the form n H2O). Consider explicitly setting the\n",
"per-element contributions if you know them, or\n",
"including a variety of structures of different\n",
"compositions in the training set.\n",
"\n",
"[graph-pes INFO]: \n",
"Number of learnable params:\n",
" base (MACEWrapper) : 3,847,696\n",
" auto_offset (LearnableOffset): 2\n",
"\n",
"[graph-pes INFO]: Sanity checking the model...\n",
"[graph-pes INFO]: Starting fit...\n",
" valid/metrics valid/metrics valid/metrics valid/metrics valid/metrics valid/metrics valid/metrics valid/metrics valid/metrics valid/metrics timer/its_per_s timer/its_per_s\n",
" epoch time per_atom_energy_rmse per_atom_energy_mae energy_rmse energy_mae forces_rmse forces_mae stress_rmse stress_mae virial_rmse virial_mae train valid\n",
" 1 7.8 0.00716 0.00673 0.77302 0.72710 0.10150 0.07997 0.00925 0.00584 128.09291 102.27287 0.80064 4.89904\n",
" 2 16.1 0.00905 0.00825 0.97704 0.89060 0.05779 0.04414 0.00524 0.00346 128.24873 103.04964 0.70522 4.23109\n",
" 3 24.6 0.00174 0.00132 0.18810 0.14233 0.03947 0.02871 0.00277 0.00183 128.95222 103.92886 0.64683 4.50949\n",
" 4 32.7 0.00432 0.00419 0.46684 0.45242 0.03426 0.02702 0.00179 0.00117 130.01169 104.82194 0.71276 4.59975\n",
" 5 40.1 0.00162 0.00142 0.17489 0.15320 0.03540 0.02763 0.00139 0.00091 130.51797 105.21088 0.78493 4.31182\n",
" 6 48.0 0.00332 0.00312 0.35861 0.33645 0.02760 0.02140 0.00186 0.00125 129.84111 104.61426 0.67797 4.33375\n",
" 7 56.0 0.00225 0.00195 0.24346 0.21077 0.02652 0.02051 0.00180 0.00125 129.79610 104.55513 0.67476 4.14495\n",
" 8 64.0 0.00140 0.00090 0.15127 0.09751 0.02467 0.01905 0.00183 0.00127 129.90742 104.63938 0.75873 3.79053\n",
" 9 71.9 0.00206 0.00177 0.22223 0.19067 0.02435 0.01856 0.00168 0.00113 130.40732 105.05786 0.75075 3.73735\n",
" 10 79.7 0.00240 0.00231 0.25874 0.24932 0.02326 0.01803 0.00110 0.00075 130.71355 105.33549 0.68776 4.25472\n",
" 11 87.3 0.00208 0.00199 0.22486 0.21538 0.02311 0.01779 0.00106 0.00070 130.88290 105.47400 0.76104 4.32935\n",
" 12 95.0 0.00122 0.00101 0.13136 0.10957 0.02261 0.01729 0.00115 0.00075 130.93184 105.50145 0.78802 4.60993\n",
" 13 101.1 0.00068 0.00040 0.07350 0.04363 0.02162 0.01665 0.00108 0.00072 130.93443 105.49496 1.02987 5.04530\n",
" 14 107.4 0.00062 0.00047 0.06713 0.05095 0.02094 0.01612 0.00106 0.00070 131.02118 105.55566 0.76511 6.38889\n",
" 15 114.5 0.00092 0.00078 0.09890 0.08396 0.02081 0.01601 0.00087 0.00058 131.20523 105.70125 0.77821 6.11329\n",
" 16 123.3 0.00220 0.00217 0.23743 0.23433 0.02170 0.01662 0.00069 0.00045 131.44841 105.88866 0.65963 3.55058\n",
" 17 132.7 0.00166 0.00158 0.17911 0.17083 0.01936 0.01502 0.00088 0.00059 131.26625 105.73106 0.54230 3.37123\n",
" 18 140.1 0.00130 0.00126 0.14066 0.13652 0.01936 0.01496 0.00062 0.00042 131.49023 105.90979 0.76336 4.44092\n",
" 19 147.6 0.00111 0.00109 0.12030 0.11782 0.02065 0.01572 0.00047 0.00031 131.74437 106.11167 0.75529 4.32136\n",
" 20 155.2 0.00304 0.00303 0.32831 0.32720 0.01846 0.01435 0.00045 0.00033 131.59532 105.98393 0.80128 5.05800\n",
"`Trainer.fit` stopped: `max_epochs=20` reached.\n",
"[graph-pes INFO]: Loading best weights from \"/Users/john/projects/graph-pes/docs/source/quickstart/graph-pes-results/mp0-fine-tune/checkpoints/best.ckpt\"\n",
"[graph-pes INFO]: Training complete.\n",
"[graph-pes INFO]: Testing best model...\n",
"GPU available: True (mps), used: False\n",
"TPU available: False, using: 0 TPU cores\n",
"HPU available: False, using: 0 HPUs\n",
"┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓\n",
"┃\u001b[1m \u001b[0m\u001b[1m Test metric \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1m DataLoader 0 \u001b[0m\u001b[1m \u001b[0m┃\n",
"┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┩\n",
"│\u001b[36m \u001b[0m\u001b[36m test/train/energy_mae \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m 0.08073119819164276 \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36m test/train/energy_rmse \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m 0.09959671646356583 \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36m test/train/forces_mae \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m 0.014654574915766716 \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36m test/train/forces_rmse \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m 0.0190340057015419 \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36mtest/train/forces_rmse_batchwise\u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m 0.018963992595672607 \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36m test/train/per_atom_energy_mae \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m 0.0007476329919882119 \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36mtest/train/per_atom_energy_rmse \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m 0.0009222375811077654 \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36m test/train/stress_mae \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m 0.00044360198080539703 \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36m test/train/stress_rmse \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m 0.0006201165379025042 \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36m test/train/virial_mae \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m 97.18195343017578 \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36m test/train/virial_rmse \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m 127.16856384277344 \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36m test/valid/energy_mae \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36m test/valid/energy_rmse \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36m test/valid/forces_mae \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36m test/valid/forces_rmse \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36mtest/valid/forces_rmse_batchwise\u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36m test/valid/per_atom_energy_mae \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36mtest/valid/per_atom_energy_rmse \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36m test/valid/stress_mae \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36m test/valid/stress_rmse \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36m test/valid/virial_mae \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36m test/valid/virial_rmse \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m \u001b[0m\u001b[35m \u001b[0m│\n",
"└──────────────────────────────────┴──────────────────────────────────┘\n",
"┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓\n",
"┃\u001b[1m \u001b[0m\u001b[1m Test metric \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1m DataLoader 1 \u001b[0m\u001b[1m \u001b[0m┃\n",
"┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┩\n",
"│\u001b[36m \u001b[0m\u001b[36m test/train/energy_mae \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36m test/train/energy_rmse \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36m test/train/forces_mae \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36m test/train/forces_rmse \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36mtest/train/forces_rmse_batchwise\u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36m test/train/per_atom_energy_mae \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36mtest/train/per_atom_energy_rmse \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36m test/train/stress_mae \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36m test/train/stress_rmse \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36m test/train/virial_mae \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36m test/train/virial_rmse \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36m test/valid/energy_mae \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m 0.13652344048023224 \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36m test/valid/energy_rmse \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m 0.10905405879020691 \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36m test/valid/forces_mae \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m 0.014955026097595692 \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36m test/valid/forces_rmse \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m 0.01909947581589222 \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36mtest/valid/forces_rmse_batchwise\u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m 0.019314901903271675 \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36m test/valid/per_atom_energy_mae \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m 0.001264381455257535 \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36mtest/valid/per_atom_energy_rmse \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m 0.0010098782368004322 \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36m test/valid/stress_mae \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m 0.00042308951378799975 \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36m test/valid/stress_rmse \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m 0.0006206676480360329 \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36m test/valid/virial_mae \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m 105.9097900390625 \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36m test/valid/virial_rmse \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m 128.0445556640625 \u001b[0m\u001b[35m \u001b[0m│\n",
"└──────────────────────────────────┴──────────────────────────────────┘\n",
"[graph-pes INFO]: Testing complete.\n",
"[graph-pes INFO]: Awaiting final Lightning and W&B shutdown...\n"
]
}
],
"source": [
"!graph-pes-train fine-tune.yaml mp0.yaml"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Nice! Fine-tuning on just 20 structures has brought the validation error down significantly 😊 Of course in real life, one would train for longer, and perhaps make use of early stopping to prevent overfitting to the small amounts of data - see the [docs](https://jla-gardner.github.io/graph-pes/cli/graph-pes-train/complete-docs.html) for the relevant config options you need to pass to do that.\n",
"\n",
"Lets load in our fine-tuned model and check that everything is working as expected:"
]
},
{
"cell_type": "code",
"execution_count": 13,
"metadata": {},
"outputs": [
{
"data": {
"image/png": "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",
"text/plain": [
""
]
},
"metadata": {
"image/png": {
"height": 329,
"width": 385
}
},
"output_type": "display_data"
},
{
"data": {
"text/plain": [
"Text(0.5, 1.0, 'Fine-tuned MP-0-small')"
]
},
"execution_count": 13,
"metadata": {},
"output_type": "execute_result"
},
{
"data": {
"image/png": "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",
"text/plain": [
""
]
},
"metadata": {
"image/png": {
"height": 329,
"width": 390
}
},
"output_type": "display_data"
}
],
"source": [
"from graph_pes.models import load_model\n",
"\n",
"test_set = dataset[50:100]\n",
"\n",
"parity_plot(\n",
" adjusted_mp0,\n",
" test_set,\n",
" property=\"energy_per_atom\",\n",
" units=\"eV/atom\",\n",
")\n",
"plt.title(\"Offset-adjusted base MP-0-small\")\n",
"plt.show()\n",
"\n",
"fine_tuned_mp0 = load_model(\"graph-pes-results/mp0-fine-tune/model.pt\").eval()\n",
"parity_plot(\n",
" fine_tuned_mp0,\n",
" test_set,\n",
" property=\"energy_per_atom\",\n",
" units=\"eV/atom\",\n",
")\n",
"plt.title(\"Fine-tuned MP-0-small\")"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Swapping out the ``MP-0`` model for the ``Orb`` model is trivial! We just need to change the model definition in the config file. Here we're being extra fancy, and freezing all the model parameters except those in the read-out head (see [graph_pes.models.freeze_all_except](https://jla-gardner.github.io/graph-pes/models/root.html#graph_pes.models.freeze_all_except) for details):"
]
},
{
"cell_type": "code",
"execution_count": 24,
"metadata": {},
"outputs": [],
"source": [
"orb_config = \"\"\"\n",
"model:\n",
" +freeze_all_except:\n",
" model:\n",
" +orb_model:\n",
" name: orb-d3-xs-v2\n",
" pattern: _orb\\\\.heads.*\n",
"\n",
"general:\n",
" run_id: orb-fine-tune\n",
"\"\"\"\n",
"\n",
"with open(\"orb.yaml\", \"w\") as f:\n",
" f.write(orb_config)"
]
},
{
"cell_type": "code",
"execution_count": 25,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"/opt/miniconda3/envs/mace/lib/python3.11/site-packages/e3nn/o3/_wigner.py:10: UserWarning: Environment variable TORCH_FORCE_NO_WEIGHTS_ONLY_LOAD detected, since the`weights_only` argument was not explicitly passed to `torch.load`, forcing weights_only=False.\n",
" _Jd, _W3j_flat, _W3j_indices = torch.load(os.path.join(os.path.dirname(__file__), 'constants.pt'))\n",
"[graph-pes INFO]: Started `graph-pes-train` at 2025-04-11 17:16:02.923\n",
"/opt/miniconda3/envs/mace/lib/python3.11/site-packages/orb_models/utils.py:30: UserWarning: Setting global torch default dtype to torch.float32.\n",
" warnings.warn(f\"Setting global torch default dtype to {torch_dtype}.\")\n",
"[graph-pes INFO]: Successfully parsed config.\n",
"[graph-pes INFO]: Logging to graph-pes-results/orb-fine-tune/rank-0.log\n",
"[graph-pes INFO]: ID for this training run: orb-fine-tune\n",
"[graph-pes INFO]: \n",
"Output for this training run can be found at:\n",
" └─ graph-pes-results/orb-fine-tune\n",
" ├─ rank-0.log # find a verbose log here\n",
" ├─ model.pt # the best model (according to valid/loss/total)\n",
" ├─ lammps_model.pt # the best model deployed to LAMMPS\n",
" ├─ train-config.yaml # the complete config used for this run\n",
" └─ summary.yaml # the summary of the training run\n",
"\n",
"GPU available: True (mps), used: False\n",
"TPU available: False, using: 0 TPU cores\n",
"HPU available: False, using: 0 HPUs\n",
"[graph-pes INFO]: Preparing data\n",
"[graph-pes INFO]: Setting up datasets\n",
"[graph-pes INFO]: Pre-fitting the model on 20 samples\n",
"[graph-pes INFO]: \n",
"Attempting to automatically detect the offset energy for each element.\n",
"We do this by first generating predictions for each training structure \n",
"(up to `config.fitting.max_n_pre_fit` if specified). \n",
"This is a slow process! If you already know the reference energies (or the\n",
"difference in reference energies if you are fine-tuning an existing model to a\n",
"different level of theory), \n",
"we recommend setting `config.fitting.auto_fit_reference_energies` to `False`\n",
"and manually specifying a `LearnableOffset` component of your model.\n",
"\n",
"See the \"Fine-tuning\" tutorial in the docs for more information: \n",
"https://jla-gardner.github.io/graph-pes/quickstart/fine-tuning.html\n",
"\n",
"[graph-pes WARNING]: \n",
"We are attempting to guess the mean per-element\n",
"contribution for a per-structure quantity (usually\n",
"the total energy). \n",
"\n",
"However, the composition of the training set is such that \n",
"no unique solution is possible. \n",
"\n",
"This is probably because you are training on structures\n",
"all with the same composition (e.g. all structures are\n",
"of the form n H2O). Consider explicitly setting the\n",
"per-element contributions if you know them, or\n",
"including a variety of structures of different\n",
"compositions in the training set.\n",
"\n",
"[graph-pes INFO]: \n",
"Number of learnable params:\n",
" base (OrbWrapper) : 200,064\n",
" auto_offset (LearnableOffset): 2\n",
"\n",
"[graph-pes INFO]: Sanity checking the model...\n",
"[graph-pes INFO]: Starting fit...\n",
" valid/metrics valid/metrics valid/metrics valid/metrics valid/metrics valid/metrics valid/metrics valid/metrics valid/metrics valid/metrics timer/its_per_s timer/its_per_s\n",
" epoch time per_atom_energy_rmse per_atom_energy_mae energy_rmse energy_mae forces_rmse forces_mae stress_rmse stress_mae virial_rmse virial_mae train valid\n",
" 1 1.8 0.01108 0.00818 1.19668 0.88313 0.16017 0.12256 0.02922 0.02326 76.28781 61.23480 3.80228 5.45729\n",
" 2 3.6 0.00770 0.00585 0.83182 0.63171 0.15506 0.11872 0.02922 0.02326 76.28781 61.23480 4.13223 11.88725\n",
" 3 5.3 0.00467 0.00438 0.50436 0.47307 0.15038 0.11539 0.02922 0.02326 76.28781 61.23480 3.86100 10.40802\n",
" 4 7.1 0.00369 0.00305 0.39796 0.32947 0.14623 0.11259 0.02922 0.02326 76.28781 61.23480 4.03226 10.95286\n",
" 5 8.9 0.00262 0.00232 0.28348 0.25066 0.14247 0.11010 0.02922 0.02326 76.28781 61.23480 3.04878 10.65872\n",
" 6 10.7 0.00185 0.00136 0.19947 0.14702 0.13893 0.10778 0.02922 0.02326 76.28781 61.23480 4.42478 11.21332\n",
" 7 12.5 0.00217 0.00192 0.23384 0.20730 0.13528 0.10521 0.02922 0.02326 76.28781 61.23480 3.90625 11.81803\n",
" 8 14.1 0.00266 0.00254 0.28749 0.27485 0.13179 0.10278 0.02922 0.02326 76.28781 61.23480 4.31034 10.62050\n",
" 9 16.0 0.00250 0.00214 0.27035 0.23108 0.12831 0.10037 0.02922 0.02326 76.28781 61.23480 3.80228 8.25962\n",
" 10 17.8 0.00269 0.00243 0.29018 0.26196 0.12483 0.09784 0.02922 0.02326 76.28781 61.23480 4.52489 10.93097\n",
" 11 19.5 0.00218 0.00198 0.23518 0.21355 0.12133 0.09523 0.02922 0.02326 76.28781 61.23480 4.40529 11.30330\n",
" 12 21.1 0.00259 0.00245 0.27953 0.26460 0.11804 0.09287 0.02922 0.02326 76.28781 61.23480 4.36681 11.35335\n",
" 13 22.8 0.00248 0.00226 0.26789 0.24412 0.11474 0.09049 0.02922 0.02326 76.28781 61.23480 4.08163 11.98733\n",
" 14 24.4 0.00215 0.00185 0.23196 0.19985 0.11158 0.08824 0.02922 0.02326 76.28781 61.23480 3.42466 11.44267\n",
" 15 26.0 0.00353 0.00302 0.38135 0.32583 0.10840 0.08593 0.02922 0.02326 76.28781 61.23480 3.41297 11.50141\n",
" 16 27.7 0.00368 0.00289 0.39733 0.31230 0.10509 0.08339 0.02922 0.02326 76.28781 61.23480 3.48432 11.62694\n",
" 17 29.4 0.00328 0.00277 0.35471 0.29954 0.10226 0.08139 0.02922 0.02326 76.28781 61.23480 4.42478 9.53071\n",
" 18 31.1 0.00258 0.00209 0.27859 0.22561 0.09923 0.07914 0.02922 0.02326 76.28781 61.23480 2.79330 11.77056\n",
" 19 32.8 0.00365 0.00313 0.39438 0.33806 0.09616 0.07678 0.02922 0.02326 76.28781 61.23480 2.94118 10.99205\n",
" 20 34.5 0.00269 0.00218 0.29059 0.23569 0.09309 0.07437 0.02922 0.02326 76.28781 61.23480 4.00000 10.98432\n",
"`Trainer.fit` stopped: `max_epochs=20` reached.\n",
"[graph-pes INFO]: Loading best weights from \"/Users/john/projects/graph-pes/docs/source/quickstart/graph-pes-results/orb-fine-tune/checkpoints/best.ckpt\"\n",
"[graph-pes INFO]: Training complete.\n",
"[graph-pes INFO]: Testing best model...\n",
"GPU available: True (mps), used: False\n",
"TPU available: False, using: 0 TPU cores\n",
"HPU available: False, using: 0 HPUs\n",
"┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓\n",
"┃\u001b[1m \u001b[0m\u001b[1m Test metric \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1m DataLoader 0 \u001b[0m\u001b[1m \u001b[0m┃\n",
"┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┩\n",
"│\u001b[36m \u001b[0m\u001b[36m test/train/energy_mae \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m 0.16972656548023224 \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36m test/train/energy_rmse \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m 0.23350417613983154 \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36m test/train/forces_mae \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m 0.07444514334201813 \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36m test/train/forces_rmse \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m 0.09575104713439941 \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36mtest/train/forces_rmse_batchwise\u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m 0.09573473781347275 \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36m test/train/per_atom_energy_mae \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m 0.0015714168548583984 \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36mtest/train/per_atom_energy_rmse \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m 0.002161899348720908 \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36m test/train/stress_mae \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m 0.02080981805920601 \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36m test/train/stress_rmse \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m 0.0265880785882473 \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36m test/train/virial_mae \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m 61.88399124145508 \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36m test/train/virial_rmse \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m 92.37788391113281 \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36m test/valid/energy_mae \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36m test/valid/energy_rmse \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36m test/valid/forces_mae \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36m test/valid/forces_rmse \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36mtest/valid/forces_rmse_batchwise\u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36m test/valid/per_atom_energy_mae \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36mtest/valid/per_atom_energy_rmse \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36m test/valid/stress_mae \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36m test/valid/stress_rmse \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36m test/valid/virial_mae \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36m test/valid/virial_rmse \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m \u001b[0m\u001b[35m \u001b[0m│\n",
"└──────────────────────────────────┴──────────────────────────────────┘\n",
"┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓\n",
"┃\u001b[1m \u001b[0m\u001b[1m Test metric \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1m DataLoader 1 \u001b[0m\u001b[1m \u001b[0m┃\n",
"┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┩\n",
"│\u001b[36m \u001b[0m\u001b[36m test/train/energy_mae \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36m test/train/energy_rmse \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36m test/train/forces_mae \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36m test/train/forces_rmse \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36mtest/train/forces_rmse_batchwise\u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36m test/train/per_atom_energy_mae \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36mtest/train/per_atom_energy_rmse \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36m test/train/stress_mae \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36m test/train/stress_rmse \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36m test/train/virial_mae \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36m test/train/virial_rmse \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36m test/valid/energy_mae \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m 0.23569336533546448 \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36m test/valid/energy_rmse \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m 0.2459844946861267 \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36m test/valid/forces_mae \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m 0.07436826825141907 \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36m test/valid/forces_rmse \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m 0.09522523730993271 \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36mtest/valid/forces_rmse_batchwise\u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m 0.09307248890399933 \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36m test/valid/per_atom_energy_mae \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m 0.002182579133659601 \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36mtest/valid/per_atom_energy_rmse \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m 0.002277533058077097 \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36m test/valid/stress_mae \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m 0.023262472823262215 \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36m test/valid/stress_rmse \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m 0.027134379372000694 \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36m test/valid/virial_mae \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m 61.23480224609375 \u001b[0m\u001b[35m \u001b[0m│\n",
"│\u001b[36m \u001b[0m\u001b[36m test/valid/virial_rmse \u001b[0m\u001b[36m \u001b[0m│\u001b[35m \u001b[0m\u001b[35m 89.39186096191406 \u001b[0m\u001b[35m \u001b[0m│\n",
"└──────────────────────────────────┴──────────────────────────────────┘\n",
"[graph-pes INFO]: Testing complete.\n",
"[graph-pes INFO]: Awaiting final Lightning and W&B shutdown...\n"
]
}
],
"source": [
"!graph-pes-train fine-tune.yaml orb.yaml"
]
},
{
"cell_type": "code",
"execution_count": 26,
"metadata": {},
"outputs": [
{
"data": {
"image/png": "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",
"text/plain": [
""
]
},
"metadata": {
"image/png": {
"height": 329,
"width": 390
}
},
"output_type": "display_data"
}
],
"source": [
"fine_tuned_orb = load_model(\"graph-pes-results/orb-fine-tune/model.pt\").eval()\n",
"parity_plot(\n",
" fine_tuned_orb,\n",
" test_set,\n",
" property=\"energy_per_atom\",\n",
" units=\"eV/atom\",\n",
")\n",
"plt.title(\"Fine-tuned orb-d3-xs-v2\");"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Again, we aren't trying to make direct comparisons between these two models here - the instances we have chosen are both small models to show proof of concept.\n",
"\n",
"If you want to explore fine-tuning various foundation models, we strongly recommend that you tune the hyperparameters of the fine-tuning process for each one separately."
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Fine-tuning other models\n",
"\n",
"Of course, you can also fine-tune other models that you have trained yourself! In this case, rather than using the foundation model interfaces, use the ``load_model`` function to load your model from disk. Remember to pass the ``auto_fit_reference_energies=true`` flag to the ``graph-pes-train`` command to ensure that this model correctly updates its reference energies:\n",
"\n",
"```yaml\n",
"model:\n",
" +load_model:\n",
" path: path/to/model.pt\n",
"\n",
"fitting:\n",
" auto_fit_reference_energies: true\n",
"\n",
"# ... <-- other config options as normal\n",
"```"
]
}
],
"metadata": {
"kernelspec": {
"display_name": "mace",
"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": "3.11.11"
}
},
"nbformat": 4,
"nbformat_minor": 2
}