{ "cells": [ { "attachments": {}, "cell_type": "markdown", "metadata": {}, "source": [ "# Training your first collective variable\n", "[![Open in Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/luigibonati/mlcolvar/blob/main/docs/notebooks/tutorials/intro_1_training.ipynb)" ] }, { "attachments": {}, "cell_type": "markdown", "metadata": {}, "source": [ "Here we will give a brief overview of the process of training a data-driven collective variable with `mlcolvar`. As example we will perform a simple regression task for a particle in a 2D toy model potential named Muller-Brown." ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [ "# Colab setup\n", "import os\n", "\n", "if os.getenv(\"COLAB_RELEASE_TAG\"):\n", " import subprocess\n", " subprocess.run('wget https://raw.githubusercontent.com/luigibonati/mlcolvar/main/colab_setup.sh', shell=True)\n", " cmd = subprocess.run('bash colab_setup.sh TUTORIAL', shell=True, stdout=subprocess.PIPE)\n", " print(cmd.stdout.decode('utf-8'))" ] }, { "cell_type": "code", "execution_count": 2, "metadata": {}, "outputs": [ { "name": "stderr", "output_type": "stream", "text": [ "/Users/luigi/opt/anaconda3/envs/pytorch13/lib/python3.9/site-packages/tqdm/auto.py:21: TqdmWarning: IProgress not found. Please update jupyter and ipywidgets. See https://ipywidgets.readthedocs.io/en/stable/user_install.html\n", " from .autonotebook import tqdm as notebook_tqdm\n" ] }, { "data": { "text/plain": [ "" ] }, "execution_count": 2, "metadata": {}, "output_type": "execute_result" }, { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "from mlcolvar.utils.plot import muller_brown_potential, plot_isolines_2D\n", "\n", "plot_isolines_2D(muller_brown_potential,levels=12,max_value=24)" ] }, { "attachments": {}, "cell_type": "markdown", "metadata": {}, "source": [ "### Outline" ] }, { "attachments": {}, "cell_type": "markdown", "metadata": {}, "source": [ "Typically, the process of constructing a data-driven CV requires the following ingredients:\n", "1) a **dataset** (with input features and possibly targets or labels)\n", "2) a **model** (e.g. a neural network) and an objective function. \n", "\n", "Once we have them, we can combine them and optimize the CV (via a **trainer** object). \n", "\n", "Finally, we can export the CV in order to be deployed in PLUMED." ] }, { "attachments": {}, "cell_type": "markdown", "metadata": {}, "source": [ "### Load files" ] }, { "attachments": {}, "cell_type": "markdown", "metadata": {}, "source": [ "We will load a COLVAR file produced by PLUMED which contains:\n", "- the descriptors (features) which will be used as input of our model, in this case x and y coordinates of the particle\n", "- the quantity to be reproduced (target) which in this case is the potential energy\n", "\n", "To do this we can use the `load_dataframe` function contained in the `mlcolvar.io` module." ] }, { "cell_type": "code", "execution_count": 5, "metadata": {}, "outputs": [ { "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", "
timep.xp.yp.zenepot.biaspot.ene_biaslwall.biaslwall.force2uwall.biasuwall.force2
00.00.5000000.0000000.06.5809816.5809816.5809810.00.00.00.0
11.00.2858030.3514470.011.50674011.50674011.5067400.00.00.00.0
22.0-0.0042930.5907100.011.82163711.82163711.8216370.00.00.00.0
33.0-0.5302080.7146880.016.81288616.81288616.8128860.00.00.00.0
44.0-1.0152360.9783060.08.8215148.8215148.8215140.00.00.00.0
\n", "
" ], "text/plain": [ " time p.x p.y p.z ene pot.bias pot.ene_bias \\\n", "0 0.0 0.500000 0.000000 0.0 6.580981 6.580981 6.580981 \n", "1 1.0 0.285803 0.351447 0.0 11.506740 11.506740 11.506740 \n", "2 2.0 -0.004293 0.590710 0.0 11.821637 11.821637 11.821637 \n", "3 3.0 -0.530208 0.714688 0.0 16.812886 16.812886 16.812886 \n", "4 4.0 -1.015236 0.978306 0.0 8.821514 8.821514 8.821514 \n", "\n", " lwall.bias lwall.force2 uwall.bias uwall.force2 \n", "0 0.0 0.0 0.0 0.0 \n", "1 0.0 0.0 0.0 0.0 \n", "2 0.0 0.0 0.0 0.0 \n", "3 0.0 0.0 0.0 0.0 \n", "4 0.0 0.0 0.0 0.0 " ] }, "execution_count": 5, "metadata": {}, "output_type": "execute_result" } ], "source": [ "from mlcolvar.io import load_dataframe\n", "\n", "filename = \"data/muller-brown/unbiased/high-temp/COLVAR\" \n", "df = load_dataframe(filename)\n", "df.head()" ] }, { "attachments": {}, "cell_type": "markdown", "metadata": {}, "source": [ "### Create dataset" ] }, { "attachments": {}, "cell_type": "markdown", "metadata": {}, "source": [ "We can now create the dataset which will be used to optimize the CV. To do this we can use the `DictDataset` implemented in the `mlcolvar.data` module. In case of supervised regression task the keys should be 'data' and 'target'." ] }, { "cell_type": "code", "execution_count": 6, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "DictDataset( \"data\": [5001, 2], \"target\": [5001, 1] )" ] }, "execution_count": 6, "metadata": {}, "output_type": "execute_result" } ], "source": [ "from mlcolvar.data import DictDataset\n", "from torch import Tensor\n", "\n", "# define input and target features\n", "X = df.filter(regex='p.x|p.y').values #x,y coordinates\n", "y = df[['ene']].values #target energy\n", "\n", "# create dataset specifying a dictionary\n", "dataset = DictDataset(dict(data=Tensor(X),target=Tensor(y)))\n", "dataset" ] }, { "attachments": {}, "cell_type": "markdown", "metadata": {}, "source": [ "In order to avoid overfitting (i.e. learning to reproduce the training dataset but without the ability to generalize to similar data), it is always good to split the training data into training and validation. \n", "\n", "Here we use the `lightning` framework which takes care of these aspects through an object called `DataModule`. This takes a dataset and creates training, validation (and optionally also testing) dataloaders which are used by pytorch in the training loop. They are responsible also for dividing the data in batches, shuffling the data and so on, but all of this happens under the hood. We implemented a `DictModule` which takes as input the dataset which we created above, as well as an indication on how to split data in train and validation sets (here we use 80/20% ratio.)" ] }, { "cell_type": "code", "execution_count": 7, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "DictModule(dataset -> DictDataset( \"data\": [5001, 2], \"target\": [5001, 1] ),\n", "\t\t train_loader -> DictLoader(length=0.8, batch_size=1024, shuffle=True),\n", "\t\t valid_loader -> DictLoader(length=0.2, batch_size=1024, shuffle=True))" ] }, "execution_count": 7, "metadata": {}, "output_type": "execute_result" } ], "source": [ "from mlcolvar.data import DictModule\n", "\n", "datamodule = DictModule(dataset,lengths=[0.8,0.2],batch_size=1024)\n", "datamodule" ] }, { "attachments": {}, "cell_type": "markdown", "metadata": {}, "source": [ "### Define the model" ] }, { "attachments": {}, "cell_type": "markdown", "metadata": {}, "source": [ "Having organized the data we can now define the model that we are going to optimize in order to accomplish our regression task. The cvs are implemented in the `mlcolvar.cvs` module. In particular we are going to use a `RegressionCV`, which is implemented as a pytorch lightning module. It is composed by two blocks which are executed one after the other: 'norm_in' and 'nn'. The first is a normalization layer which standardize the data based on statistics on the training set, the second the actual FeedForward neural network. These blocks can be customized by passing a dictionary of keyword arguments as described below.\n", "\n", "Here we will use a feedforward neural network with three layers and [25,50,25] nodes per layer." ] }, { "cell_type": "code", "execution_count": 8, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "RegressionCV(\n", " (norm_in): Normalization(in_features=2, out_features=2, mode=mean_std)\n", " (nn): FeedForward(\n", " (nn): Sequential(\n", " (0): Linear(in_features=2, out_features=25, bias=True)\n", " (1): ReLU(inplace=True)\n", " (2): Linear(in_features=25, out_features=50, bias=True)\n", " (3): ReLU(inplace=True)\n", " (4): Linear(in_features=50, out_features=25, bias=True)\n", " (5): ReLU(inplace=True)\n", " (6): Linear(in_features=25, out_features=1, bias=True)\n", " )\n", " )\n", ")" ] }, "execution_count": 8, "metadata": {}, "output_type": "execute_result" } ], "source": [ "from mlcolvar.cvs import RegressionCV\n", "\n", "# specify number of nodes per layer (including input and output dimensions) \n", "layers = [X.shape[1],25,50,25,1]\n", "\n", "# specify options for the nn block\n", "nn_args = {'activation': 'relu'}\n", "# specify options for the normalization layer (can be disabled by setting the norm args equal to None or False)\n", "norm_args = {}\n", "\n", "model = RegressionCV (layers, options={'norm_in':norm_args , 'nn': nn_args} )\n", "model" ] }, { "attachments": {}, "cell_type": "markdown", "metadata": {}, "source": [ "The loss function is embedded inside the CV object, for a `RegressionCV` the default is the Mean Square Error (MSE), you will see how to customize it in the following tutorials." ] }, { "attachments": {}, "cell_type": "markdown", "metadata": {}, "source": [ "### Define the trainer" ] }, { "attachments": {}, "cell_type": "markdown", "metadata": {}, "source": [ "The last ingredient that we need to define is a lightning `Trainer`, which ties together the module and the datamodule and performs the training/validation loop. This can be customized with all Lightning features, from logging the metrics to TensorBoard or using EarlyStopping or Learning Rate schedulers. In addition, it can be used to specify the devices to be used for training the CVs, without any extra modification to the code.\n", "\n", "In this simple example we will only use a `MetricsCallback` to save the metrics into a dictionary as well as `EarlyStopping` criterion acting on the \"valid_loss\" (it will automatically stop the training after the validation loss does not increase for `patience` epochs.)" ] }, { "cell_type": "code", "execution_count": 127, "metadata": {}, "outputs": [ { "name": "stderr", "output_type": "stream", "text": [ "GPU available: False, used: False\n", "TPU available: False, using: 0 TPU cores\n", "IPU available: False, using: 0 IPUs\n", "HPU available: False, using: 0 HPUs\n" ] } ], "source": [ "from lightning import Trainer\n", "from lightning.pytorch.callbacks.early_stopping import EarlyStopping\n", "from mlcolvar.utils.trainer import MetricsCallback\n", "\n", "# define callbacks\n", "metrics = MetricsCallback()\n", "early_stopping = EarlyStopping(monitor=\"valid_loss\", patience=10, min_delta=1e-5)\n", "\n", "# define trainer\n", "trainer = Trainer(callbacks=[metrics, early_stopping],\n", " logger=None, enable_checkpointing=False) #disable logger and checkpointing in this tutorial" ] }, { "attachments": {}, "cell_type": "markdown", "metadata": {}, "source": [ "### Optimize it!" ] }, { "attachments": {}, "cell_type": "markdown", "metadata": {}, "source": [ "Finally we can call the fit method of the trainer and optimize our model!" ] }, { "cell_type": "code", "execution_count": 128, "metadata": {}, "outputs": [ { "name": "stderr", "output_type": "stream", "text": [ "/Users/luigi/opt/anaconda3/envs/pytorch13/lib/python3.9/site-packages/lightning/loops/utilities.py:70: PossibleUserWarning: `max_epochs` was not set. Setting it to 1000 epochs. To train without an epoch limit, set `max_epochs=-1`.\n", " rank_zero_warn(\n", "\n", " | Name | Type | Params | In sizes | Out sizes\n", "----------------------------------------------------------------\n", "0 | norm_in | Normalization | 0 | [2] | [2] \n", "1 | nn | FeedForward | 2.7 K | [2] | [1] \n", "----------------------------------------------------------------\n", "2.7 K Trainable params\n", "0 Non-trainable params\n", "2.7 K Total params\n", "0.011 Total estimated model params size (MB)\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ " " ] }, { "name": "stderr", "output_type": "stream", "text": [ "/Users/luigi/opt/anaconda3/envs/pytorch13/lib/python3.9/site-packages/lightning/loops/fit_loop.py:280: PossibleUserWarning: The number of training batches (4) is smaller than the logging interval Trainer(log_every_n_steps=50). Set a lower value for log_every_n_steps if you want to see logs for the training epoch.\n", " rank_zero_warn(\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "Epoch 471: 100%|██████████| 4/4 [00:00<00:00, 46.14it/s, v_num=8] \n" ] } ], "source": [ "trainer.fit( model, datamodule )" ] }, { "attachments": {}, "cell_type": "markdown", "metadata": {}, "source": [ "The train/valid losses are saved into the metrics object. We can use an helper function to plot them:" ] }, { "cell_type": "code", "execution_count": 129, "metadata": {}, "outputs": [ { "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAAAdUAAAGHCAYAAADm/8gLAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjcuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/bCgiHAAAACXBIWXMAAA9hAAAPYQGoP6dpAABWQUlEQVR4nO3dd3wUdf7H8dembXoIhCQkhIQSAqGErgEFKdIElKIeooLlPBTbYbmzi+JPRQU8aVbsHVBUFEFpAgICoRcRQgJJgBBSSc/8/lhciYQQwpJJeT8fj31c9juzs5+ZeHkz35n5fi2GYRiIiIjIBXMyuwAREZHaQqEqIiLiIApVERERB1GoioiIOIhCVURExEEUqiIiIg6iUBUREXEQhaqIiIiDKFRFREQcRKEqcg7vvvsuFouF3377zexSztsVV1zBFVdcYXYZInWGi9kFiMjFM2vWLLNLEKlTFKoiNYRhGOTl5eHh4VHhz0RHR1/EisxVWFiIxWLBxUV/xqT6UPeviIP8/vvv3HDDDQQGBmK1WmndujUzZ84stU5eXh4PPPAAHTp0wM/Pj/r16xMbG8vXX399xvYsFgt33303c+bMoXXr1litVt577z17d/SyZcu48847CQgIoEGDBowYMYKkpKRS2/h79298fDwWi4WXX36ZqVOn0rRpU7y9vYmNjeXXX389o4Y333yTli1bYrVaiY6O5uOPP2bcuHFERERU6Jh8/PHHxMbG4u3tjbe3Nx06dODtt9+2L4+IiGDcuHFnfO7vdS9fvhyLxcIHH3zAAw88QGhoKFarlR07dmCxWEpt80/ff/89FouFhQsX2tsq8jsqKSlh8uTJREVF4eHhQb169Wjfvj2vvvpqhfZZ6jb9E0/EAXbu3En37t1p0qQJr7zyCsHBwSxevJh7772X1NRUnnrqKQDy8/NJS0vjwQcfJDQ0lIKCApYuXcqIESOYO3cuN998c6ntfvXVV6xatYonn3yS4OBgAgMD2bBhAwC33347V111FR9//DGJiYk89NBD3Hjjjfz888/nrHfmzJm0atWK6dOnA/DEE08wePBgDhw4gJ+fHwBvvPEG//rXvxg5ciTTpk0jIyODSZMmkZ+fX6Fj8uSTT/Lss88yYsQIHnjgAfz8/Ni+fTsHDx6s6GE9wyOPPEJsbCxz5szBycmJsLAwOnbsyNy5c7nttttKrfvuu+8SGBjI4MGDgYr/jqZMmcLTTz/N448/Ts+ePSksLGT37t2kp6dXum6pQwwRKdfcuXMNwNiwYcNZ1xkwYIDRuHFjIyMjo1T73Xffbbi7uxtpaWllfq6oqMgoLCw0brvtNqNjx46llgGGn5/fGZ/9s5677rqrVPuUKVMMwEhOTra39erVy+jVq5f9/YEDBwzAaNeunVFUVGRvX79+vQEYn3zyiWEYhlFcXGwEBwcbl1xySanvOHjwoOHq6mqEh4ef9VgYhmHs37/fcHZ2NsaMGVPueuHh4cbYsWPPaP973cuWLTMAo2fPnmes+7///c8AjD179tjb0tLSDKvVajzwwAP2tor+joYMGWJ06NCh3LpFzkbdvyIXKC8vj59++onhw4fj6elJUVGR/TV48GDy8vJKda1+8cUX9OjRA29vb1xcXHB1deXtt99m165dZ2y7T58++Pv7l/m9w4YNK/W+ffv2ABU6E7zqqqtwdnY+62f37NlDSkoK1113XanPNWnShB49epxz+0uWLKG4uJgJEyacc93zMXLkyDPaxowZg9Vq5d1337W3ffLJJ+Tn53PLLbcA5/c76tatG1u2bOGuu+5i8eLFZGZmOnQfpHZTqIpcoOPHj1NUVMRrr72Gq6trqdefXY+pqakAzJ8/n+uuu47Q0FA+/PBD1q5dy4YNG7j11lvJy8s7Y9uNGjU66/c2aNCg1Hur1QpAbm7uOWs+12ePHz8OQFBQ0BmfLavt744dOwZA48aNz7nu+SjreNSvX59hw4bx/vvvU1xcDNi6frt160abNm2A8/sdPfLII7z88sv8+uuvDBo0iAYNGtC3b98a+UiVVD1dUxW5QP7+/jg7O3PTTTed9cysadOmAHz44Yc0bdqUzz77DIvFYl9+tuuUp69Tlf4M3SNHjpyxLCUl5Zyfb9iwIQCHDh0iLCzsrOu5u7uXue+pqakEBASc0X6243HLLbfwxRdfsGTJEpo0acKGDRuYPXu2ffn5/I5cXFyYOHEiEydOJD09naVLl/Loo48yYMAAEhMT8fT0PPuOS52nUBW5QJ6envTu3ZvNmzfTvn173NzczrquxWLBzc2tVDikpKSUefevmaKioggODubzzz9n4sSJ9vaEhATWrFlDSEhIuZ/v378/zs7OzJ49m9jY2LOuFxERwdatW0u17d27lz179pQZquV9X2hoKHPnzqVJkya4u7szevRo+/Lz+R2drl69eowaNYrDhw9z//33Ex8fX6sfU5ILp1AVqaCff/6Z+Pj4M9oHDx7Mq6++ymWXXcbll1/OnXfeSUREBFlZWezbt49vvvnGfkfukCFDmD9/PnfddRejRo0iMTGRZ599lkaNGvH7779X8R6dnZOTE5MmTeJf//oXo0aN4tZbbyU9PZ1JkybRqFEjnJzKv3IUERHBo48+yrPPPktubi6jR4/Gz8+PnTt3kpqayqRJkwC46aabuPHGG7nrrrsYOXIkBw8eZMqUKfYz3Ypydnbm5ptvZurUqfj6+jJixAj7Xcx/qujvaOjQobRt25YuXbrQsGFDDh48yPTp0wkPDycyMvK86pK6R6EqUkH/+c9/ymw/cOAA0dHRbNq0iWeffZbHH3+co0ePUq9ePSIjI+3X7MDWTXn06FHmzJnDO++8Q7Nmzfjvf//LoUOH7EFTXdxxxx1YLBamTJnC8OHDiYiI4L///S9ff/01CQkJ5/z8M888Q2RkJK+99hpjxozBxcWFyMhI7r33Xvs6N9xwA0lJScyZM4e5c+fStm1bZs+eXaljccstt/D8889z7Ngx+w1Kp6vo76h3797MmzePt956i8zMTIKDg7nyyit54okncHV1Pe+6pG6xGIZhmF2EiNQM6enptGzZkmuuuYY33njD7HJEqh2dqYpImVJSUnjuuefo3bs3DRo04ODBg0ybNo2srCzuu+8+s8sTqZYUqiJSJqvVSnx8PHfddRdpaWl4enpy6aWXMmfOHPujKiJSmrp/RUREHESDP4iIiDiIQlVERMRBFKoiIiIOohuVylFSUkJSUhI+Pj6mDRcnIiLmMwyDrKwsQkJCyh38RKFajqSkpHLHLRURkbolMTGx3IkiFKrl8PHxAWwH0dfX1+RqRETELJmZmYSFhdlz4WwUquX4s8vX19dXoSoiIue8FKgblURERBxEoSoiIuIgClUREREHUaiKiIg4iEJVRETEQRSqIiIiDqJQFRERcRCFqoiIiIMoVEVERBxEoVqGmTNnEh0dTdeuXR2zweP74OAqx2xLRESqLYVqGSZMmMDOnTvZsGHDhW8sKwlWPIOxbgY5+xWsIvKXiIgIpk+f7pBtLV++HIvFQnp6ukO2V5PFx8djsViIi4ur8u/W2L8Xm3cjjtVvT8Oj63HbMJNsVy+8wzqZXZWIVNIVV1xBhw4dHBKGGzZswMvL68KLkmpDZ6oXm8VCRtuxrCgMwtVSQvHa6RTmpJldlYhcJIZhUFRUVKF1GzZsiKen50WuSKqSQrUKtAhoSFj/x/ij2Bs/8vh9+WtmlyRSbeUWFp73q6ikxP75opIScgsLyftbsJ3ts+dj3LhxrFixgldffRWLxYLFYuHdd9/FYrGwePFiunTpgtVqZdWqVfzxxx9cffXVBAUF4e3tTdeuXVm6dGmp7f29+9disfDWW28xfPhwPD09iYyMZOHChed/EE+ZN28ebdq0wWq1EhERwSuvvFJq+axZs4iMjMTd3Z2goCBGjRplX/bll1/Srl07PDw8aNCgAf369SMnJ6dC3zt37lxat26Nu7s7rVq1YtasWfZlf3bNfvrpp3Tv3h13d3fatGnD8uXLS21jxYoVdOvWDavVSqNGjfjvf/9b6h8rJSUlvPjii7Ro0QKr1UqTJk147rnnSm1j//799O7dG09PT2JiYli7dm1FD12lqfu3ijRrGMz61jfSfO8cWmXv4HDiDkLD2phdlki1EzvnjfP+zJSBA+gf2QKAn//Yz8M/LKZzaAhvjxhuX2fwu+9zIi/vjM/G3TOhwt/z6quvsnfvXtq2bcszzzwDwI4dOwB4+OGHefnll2nWrBn16tXj0KFDDB48mMmTJ+Pu7s57773H0KFD2bNnD02aNDnrd0yaNIkpU6bw0ksv8dprrzFmzBgOHjxI/fr1K1wnwMaNG7nuuut4+umnuf7661mzZg133XUXDRo0YNy4cfz222/ce++9fPDBB3Tv3p20tDRWrbLd95GcnMzo0aOZMmUKw4cPJysri1WrVmEYxjm/98033+Spp55ixowZdOzYkc2bN/PPf/4TLy8vxo4da1/voYceYvr06URHRzN16lSGDRvGgQMHaNCgAYcPH2bw4MGMGzeO999/n927d/PPf/4Td3d3nn76aQAeeeQR3nzzTaZNm8Zll11GcnIyu3fvLlXLY489xssvv0xkZCSPPfYYo0ePZt++fbi4XLzoU6hWoW4d+rBl3yJiShJI3vAxoWHPnftDIlJt+Pn54ebmhqenJ8HBwQD2P+TPPPMMV155pX3dBg0aEBMTY38/efJkFixYwMKFC7n77rvP+h3jxo1j9OjRAPzf//0fr732GuvXr2fgwIHnVevUqVPp27cvTzzxBAAtW7Zk586dvPTSS4wbN46EhAS8vLwYMmQIPj4+hIeH07FjR8AWqkVFRYwYMYLw8HAA2rVrV6HvffbZZ3nllVcYMWIEAE2bNmXnzp28/vrrpUL17rvvZuTIkQDMnj2bH374gbfffpuHH36YWbNmERYWxowZM7BYLLRq1YqkpCT+85//8OSTT5KTk8Orr77KjBkz7Nts3rw5l112WalaHnzwQa666irA9o+VNm3asG/fPlq1anVex/J8KFSrWP2O18PGl2hXuI/DqUmEBoSYXZJItbJ2/B3n/RlXZ2f7z32aN2Pt+DvOmEx60bibL7i28nTp0qXU+5ycHCZNmsS3335LUlISRUVF5ObmkpCQUO522rdvb//Zy8sLHx8fjh49et717Nq1i6uvvrpUW48ePZg+fTrFxcVceeWVhIeH06xZMwYOHMjAgQPt3c4xMTH07duXdu3aMWDAAPr378+oUaPw9/cv9zuPHTtGYmIit912G//85z/t7UVFRfj5+ZVaNzY21v6zi4sLXbp0YdeuXfbaY2NjS/0Oe/ToQXZ2NocOHSIlJYX8/Hz69u1bbj2nH8tGjRoBcPTo0YsaqrqmWsXCmnXhkKUeVovB7o1fmV2OSLXj4ep63i8Xp7/+lLk4OeHh6or737r4zvZZR/n7XbwPPfQQ8+bN47nnnmPVqlXExcXRrl07CgoKyt2O699qslgslJx2zbiiDMM44x8Wp3ff+vj4sGnTJj755BMaNWrEk08+SUxMDOnp6Tg7O7NkyRK+//57oqOjee2114iKiuLAgQPlfuefdb755pvExcXZX9u3b+fXX389Z81/1lte7RaLBQ8Pj3MfAEofyz+3V5ljeT4UqlXNYiE3tDsAAWlbKb7Iv2ARcSw3NzeKi4vPud6qVasYN24cw4cPp127dgQHBxMfH3/xCzwlOjqaX375pVTbmjVraNmyJc6nzuxdXFzo168fU6ZMYevWrcTHx/Pzzz8DthDq0aMHkyZNYvPmzbi5ubFgwYJyvzMoKIjQ0FD2799PixYtSr2aNm1aat3TQ7aoqIiNGzfazyCjo6NZs2ZNqX8ErFmzBh8fH0JDQ4mMjMTDw4Offvqp8gfoIlH3rwnC2w6AQ4to65TGlvjf6dQsyuySRKSCIiIiWLduHfHx8Xh7e5/1zKdFixbMnz+foUOHYrFYeOKJJy76WdLpHnjgAbp27cqzzz7L9ddfz9q1a5kxY4b9Ttxvv/2W/fv307NnT/z9/Vm0aBElJSVERUWxbt06fvrpJ/r3709gYCDr1q3j2LFjtG7d+pzf+/TTT3Pvvffi6+vLoEGDyM/P57fffuPEiRNMnDjRvt7MmTOJjIykdevWTJs2jRMnTnDrrbcCcNdddzF9+nTuuece7r77bvbs2cNTTz3FxIkTcXJywt3dnf/85z88/PDDuLm50aNHD44dO8aOHTu47bbbLs4BrShDziojI8MAjIyMDIdvO+nL8Ybx2bXGNz+84fBti8jFs2fPHuPSSy81PDw8DMCYO3euARgnTpwotd6BAweM3r17Gx4eHkZYWJgxY8YMo1evXsZ9991nXyc8PNyYNm2a/T1gLFiwoNR2/Pz8jLlz556zrmXLlp1Rx5dffmlER0cbrq6uRpMmTYyXXnrJvmzVqlVGr169DH9/f8PDw8No37698dlnnxmGYRg7d+40BgwYYDRs2NCwWq1Gy5Ytjddee62ih8j46KOPjA4dOhhubm6Gv7+/0bNnT2P+/Pn24wIYH3/8sXHJJZcYbm5uRuvWrY2ffvqp1DaWL19udO3a1XBzczOCg4ON//znP0ZhYaF9eXFxsTF58mQjPDzcvn//93//V+o7Nm/ebF//xIkTBmAsW7aswvtxuormgcUwKnCPdB2VmZmJn58fGRkZ+Pr6OnTbB1fMJPzICpYWhdF39MtnXD8QEamN4uPjadq0KZs3b6ZDhw5ml1NhFc0DXVM1SVCzSwDo4JJGznk+gC4iItWTQtUk7o3aYlicCSAH78IMs8sRkWpu/PjxeHt7l/kaP358ldVxthq8vb3tg0fUZer+LcfF7P4FYMl/4cR+iP03hMWee30RqbOOHj1KZmZmmct8fX0JDAyskjr27dt31mWhoaEVftylpqloHujuXzPVbwEn9lN4bC+uClURKUdgYGCVBWd5WrRoYXYJ1Zq6f02U7xsBwNadv3DyHA+Ei4hI9adQNZG1YUsAWrpk88fx4yZXIyIiF0qhaiafEAyLEz5OxbTzczO7GhERuUAKVTM5u2DxOTWgfkaiubWIiMgFU6iWYebMmURHR9O1a9eL/2V+Ybb/zSh/5goREan+FKplmDBhAjt37mTDhg0X/bvyvGzTEW3Yua5CEwCLSM0XERHB9OnT7e8tFgtfffXVWdePj4/HYrEQFxd3zm0vX74ci8VCenr6Bdcp50+P1JjM+VT3r/PJI6SePEnDv00fJSK1X3Jy8jnnKpWaQWeqJnP1CwUgzDmPA2knTK5GRMwQHByM1Wo1uwxxAIWq2byDAWjoXMChtKMmFyNiIsOAojxzXudx6eX1118nNDT0jGnchg0bxtixY/njjz+4+uqrCQoKwtvbm65du7J06dJyt/n37t/169fTsWNH3N3d6dKlC5s3bz6vQ/l38+bNo02bNlitViIiInjllVdKLZ81axaRkZG4u7sTFBTEqFGj7Mu+/PJL2rVrh4eHBw0aNKBfv37k5ORcUD21mbp/zebmxUmLO55GHifTEoBOZlckYo7ifJh/sznfPeJ9cHGv0KrXXnst9957L8uWLaNv374AnDhxgsWLF/PNN9+QnZ3N4MGDmTx5Mu7u7rz33nsMHTqUPXv20KRJk3NuPycnhyFDhtCnTx8+/PBDDhw4wH333VfpXdu4cSPXXXcdTz/9NNdffz1r1qzhrrvuokGDBowbN47ffvuNe++9lw8++IDu3buTlpZmH8M3OTmZ0aNHM2XKFIYPH05WVharVq3S/R/lUKhWAznWBnjmHaYkK9nsUkTkHOrXr8/AgQP5+OOP7aH6xRdfUL9+ffr27YuzszMxMTH29SdPnsyCBQtYuHAhd9999zm3/9FHH1FcXMw777yDp6cnbdq04dChQ9x5552Vqnfq1Kn07duXJ554AoCWLVuyc+dOXnrpJcaNG0dCQgJeXl4MGTIEHx8fwsPD6dixI2AL1aKiIkaMGEF4eDgA7dq1q1QddYVCtRowPAMg7zCcTDW7FBHzOFttZ4xmffd5GDNmDHfccQezZs3CarXy0Ucf8Y9//ANnZ2dycnKYNGkS3377LUlJSRQVFZGbm0tCQsUem9u1axcxMTF4enra22JjKz82+K5du7j66qtLtfXo0YPp06dTXFzMlVdeSXh4OM2aNWPgwIEMHDiQ4cOH4+npSUxMDH379qVdu3YMGDCA/v37M2rUKN1UVQ5dU60GXH2CAHAvyFC3itRdFoutC9aMl8VyXqUOHTqUkpISvvvuOxITE1m1ahU33ngjAA899BDz5s3jueeeY9WqVcTFxdGuXTsKKji+t6P/BhiGgeVv+3f6d/j4+LBp0yY++eQTGjVqxJNPPklMTAzp6ek4OzuzZMkSvv/+e6Kjo3nttdeIioriwIEDDq2xNlGoVgPe9WyP1TTgJBl5+SZXIyLn4uHhwYgRI/joo4/45JNPaNmyJZ07dwZg1apVjBs3juHDh9OuXTuCg4OJj4+v8Lajo6PZsmULubm59rZff/210rVGR0fzyy+/lGpbs2YNLVu2xNnZGQAXFxf69evHlClT2Lp1K/Hx8fz888+A7SaqHj16MGnSJDZv3oybmxsLFiyodD21nbp/qwFXH9t0TsHOeaRkZ1HPo2I3TIiIecaMGcPQoUPZsWOH/SwVbFOjzZ8/n6FDh2KxWHjiiSfOuFO4PDfccAOPPfYYt912G48//jjx8fG8/PLLla7zgQceoGvXrjz77LNcf/31rF27lhkzZjBr1iwAvv32W/bv30/Pnj3x9/dn0aJFlJSUEBUVxbp16/jpp5/o378/gYGBrFu3jmPHjtG6detK11Pb6Uy1OvAIACDYKZ/krCyTixGRiujTpw/169dnz5493HDDDfb2adOm4e/vT/fu3Rk6dCgDBgygU6eK39Xv7e3NN998w86dO+nYsSOPPfYYL774YqXr7NSpE59//jmffvopbdu25cknn+SZZ55h3LhxANSrV4/58+fTp08fWrduzZw5c/jkk09o06YNvr6+rFy5ksGDB9OyZUsef/xxXnnlFQYNGlTpemo7i6GLeGdV0ZneL1h+Nnx9KwBfRP6Hazt2vnjfJSIi562ieaAz1erAzYsCiysAOZkpJhcjIiKVpVCtDiwWCqz1AGjj42xuLSJSrY0fPx5vb+8yX+PHjze7vDpPNypVE971QiDlGF39Nf6niJzdM888w4MPPljmsot6mUoqRKFaXXg0sP2vBoAQkXIEBgYSGBhodhlyFur+rS48bXcA56QnU3wet9+LiEj1oVCtJkpOherW/Ts4phkgRERqJIVqNeHkZQvVRs75pJ02koqIiNQcuqZaXZy6phpuLcbSsKHJxYiISGXoTLW6OBWqluJ8KFT3r4hITaRQrS5c3MDqY/v55HFzaxERkUpRqFYjJ11sz5gt+G2VyZWIiEhlKFSrkVxXW6geTz1kciUiIlIZCtVqxMnLdoOStSDd3EJERKRSFKrViLtvMAD+RjZ5RUUmVyMiIudLoVqNuNcLBSDMOY8j2dkmVyMiIuerToTq8OHD8ff3Z9SoUWaXUi6LdyMAwlxySdFk5SIiNU6dCNV7772X999/3+wyzs07CID6ToWkpB01uRgRETlfdSJUe/fujY+Pj9llnJurBzlOHgBkHD9ocjEiInK+qn2orly5kqFDhxISEoLFYuGrr746Y51Zs2bRtGlT3N3d6dy5M6tW1dznPHOttjGAi9MTTa5ERETOV7UP1ZycHGJiYpgxY0aZyz/77DPuv/9+HnvsMTZv3szll1/OoEGDSEhIqOJKHaPEvzkAwXk1s34Rkbqs2ofqoEGDmDx5MiNGjChz+dSpU7ntttu4/fbbad26NdOnTycsLIzZs2ef93fl5+eTmZlZ6lXVvMI6A3CF82Gyk7ZDiR6tERGpKap9qJanoKCAjRs30r9//1Lt/fv3Z82aNee9veeffx4/Pz/7KywszFGlVphXSAz5hjMelhK8f3kGfnoc8qs+3EVE5PzV6FBNTU2luLiYoKCgUu1BQUGkpKTY3w8YMIBrr72WRYsW0bhxYzZs2FDm9h555BEyMjLsr8REE65rurqzPuxa/ijytL0/sR+2fFj1dYiIyHmrFfOpWiyWUu8NwyjVtnjx4gptx2q1YrVaHVpbZXTrNoxnc3z5R4gbbXfMgPjl0KwPBLQyuzQRESlHjT5TDQgIwNnZudRZKcDRo0fPOHutSawuLky+sh9t2/SEpn0AyF09XVPCiYhUczU6VN3c3OjcuTNLliwp1b5kyRK6d+9uUlWOldv6Oo4Ynnjkp5G39Ak4mWp2SSIichbVPlSzs7OJi4sjLi4OgAMHDhAXF2d/ZGbixIm89dZbvPPOO+zatYt///vfJCQkMH78+Ep/58yZM4mOjqZr166O2IUL4uFdn4/9h5NU7IF7Xir8/ARkJpldloiIlMFiGIZhdhHlWb58Ob179z6jfezYsbz77ruAbfCHKVOmkJycTNu2bZk2bRo9e/a84O/OzMzEz8+PjIwMfH19L3h7lZVTUMCRo/E02zYLspLA6gs9HwX/ZqbVJCJSl1Q0D6p9qJqpuoSqXV4mrHoOThyg2Nkd516P6uYlEZEqUNE8qPbdv3Iad1/2tp1AXGE9nIvzKFk+GY5sNbsqERE5RaFawzRpGMJbHv1Zne+PU0kBJStfgMNlP3crIiJVS6Faw7i7uPDS0Kt5z6MfS/MCcDKKMNa8Agd/Mbs0EZE6T6Fahup0929ZPFxdmTbsaj507803uUFYjBKMda/BH0vNLk1EpE7TjUrlqHY3Kv1NVn4+d371NUNz13C956nHbGJugqih5hYmIlLL6EalOsDHamXW1cNY4H4p7+ScGvx/ywew4wtzCxMRqaMUqjWcr7s7s6++mu+tXflfVlNb444vYOd8cwsTEamDFKq1gL+HB69fM4zl1himZp0aEGL7p7B3kbmFiYjUMQrVWqK+pyezrxnGD5ZWzMkOtzXGvQv7fza1LhGRukShWosEeXvz2tAhfFDQguWubW2Nv70OCavNLUxEpI5QqJahuj9SU56ohgF8cN219Lz6cWjWDzBg3QxI2mR2aSIitZ4eqSlHdX+k5pyMEox1M7EkrAJnK/R+Guo3N7sqEZEaR4/UCIUlBk9mtGRtvj8U58OqFyD7qNlliYjUWgrVWszFyQmLswv/zWpLujUI8jNg1fNQkG12aSIitZK6f8tR47t/gYLiYvYdP060twv89BjkHoeGraHn4+DsanZ5IiI1grp/BQA3Z2eiAwPBsz5c/giGiwcc2wUbZoFRYnZ5IiK1ikK1Dvm9yIsnT3agBCfbYzbbPjW7JBGRWkWhWoaa/EhNeVYcOMA3J9yYnBVla9j9lWa2ERFxIF1TLUdtuKZ6uhLD4L5vv2NV/EEerp/MDa57weIEl/0HGnU0uzwRkWpL11TlDE4WC89d2Y8QXx+mpAWzzrm57brq2mlw4oDZ5YmI1HgK1TrG192dVwYNws3ZhQlJISR5hENRnu1Rm+wUs8sTEanRFKp1UOvAhjzc83KKcGJ0QmNOeoZAXjqseA5yT5hdnohIjaVQraNGtonmqqiWZJS4cMuxKIo9G0LOEVj5nAaHEBGpJIVqHWWxWHi89xU0q+/PnuwSniiIxXCvBxkJ8MuLUJRvdokiIjWOQrUO83B15eVBA/FwdWHR4Qw+9h0Crl6Qusc2TnBhrtkliojUKArVOq5Z/fo81acPAC9tS2R1kxvA1QOO7YAVk9UVLCJyHhSqwsCWkdzauRMAnyblY/R8Aty8Ie13WD7JdhOTiIickwZ/KMPMmTOZOXMmxcXF7N27t9YM/lCeEsNg/o6dXN26Fa7OzpCeACuetc1s41Efuj8ADSLNLlNExBQVHfxBoVqO2jai0vkqyjiMy5qXIeswOLlAx1uheT+zyxIRqXIaUUkqrbC4mOeWr+DxtTso6TsZQrtBSRFsfAN+/Z+us4qInIVCVc6wN/U4C3bs5Md9+9h2PMvW9dtutG2c4IRfYPGDcGSr2WWKiFQ76v4tR13u/v1uzx5cnZzpH9nir8bjv8O6GZCdbHsfcQW0uwE86plRoohIldE1VQeoy6H6d8dPnsTfwwOn4nzY8iH88aNtgYsHRI+EyMHg7GJukSIiF4muqYrDpObkcPMX83jsxyUUWlyh8+3QdzL4N4eiXNj6Ifz4ACRvMrtUERFTKVTlnHYcPcqR7Gy+3/s79377HZl5edCgJfR7DrreCe5+kJVsG4Vp5fOQedjskkVETKHu33Ko+/cvqw8e5IFFP5BXVEQjH2+eH9CfDo0a2RYWnoSd8+H376CkGJycIXoUtLra9iiOiEgNp2uqDqBQLW33sWM89P1iEjMycLZYuOvSS7ilcyecLBbbClnJEPfeX93A/s2g213g18S8okVEHECh6gAK1TNlFxTw3LLlfL/3dwAuCWvMc1f2I8DLy7aCYdgeu9n8DhTk2M5U21wHUcPASVcbRKRmUqhegLo4TOH5MAyDr3ft4oUVq8grKsLXamXiZT24unUrLH+eteaesA0WkbTR9j6wHVxyN3j4m1e4iEglKVQdQGeq5duflsajPy5h97FUALo1DuXx3lfQpF492wqGAfHLYdM7UJwPVj+IvR8C25hUsYhI5ShUHUChem5FJSV8GBfH7F/Xk19cjNXZmc9GX0+E/2lnpJmHYe10yDhoG5Up5ibbc61/ntWKiFRzek5VqoSLkxPjOnXiyzGjuaRxY7qFNSb8zzPVP/mG2p5rDb8cjBLbzUzrXoPiQlNqFhG5WHSmWg6dqZ4fwzDILSzE080NsN3UtOLAAQa3bGm71moY8Pv3sOV9W7gGx0D3B8HFanLlIiLl05mqVDmLxWIPVMMwmPTTzzz241JeXbP2zxWg5WDo+Sg4WyFlC6x6HgpzTaxaRMRxFKpy0UQHBmJ1dqZP82alFwS1h56P2cYNPrYTVkzWdHIiUiuo+7cc6v69cKk5OX89wwokZ2XRyMfH9iZtH6x8zvY8a/3mcMUkcHEzqVIRkbNT969UC6cH6t7UVIZ/+DEvrlhJYXEx1G8BVzwNbj6Q9gf8Nsd23VVEpIZSqEqV2XDoMHlFRXyydRv3f7uI3MJCqBcO3SeCxdk2EtOehWaXKSJSaQpVqTJjOsQw/arBuLu4sDohgfFfL7TNeBPYBjqOs6209WNNISciNZZCVarUFc2aMueaYfhYrWxJTuHW+QtIzcmB5v2hWT/AgF9f1fRxIlIjKVSlynVo1Ih3RgynoZcn+46ncefX35CRnw8db4WAVrZHbNZOhaICs0sVETkvClUxRWRAA94+Fay/Hz/OPd98R76B7fqq1Q8yEmHrh2aXKSJyXhSqZZg5cybR0dF07drV7FJqtSb16jH76mH4Wq1sTUnhmZ+XYVj9bHOwAuz7QddXRaRG0XOq5dBzqlVjfeIh7vx6IcWGwb2xl3Jrl86wea5tSEOrHwx8Baw6/iJiHj2nKjVGt7DG/KdXTwBeW/sry/bvh/ZjwLcx5GfYpo4TEakBFKpSLVzXri3Xt2uHASzctRuc3WzdwBYnSFwDh9aZXaKIyDm5mF2AyJ8e6nkZbYICuSqqpa2hfgtodTXsWgAb37I9z+rmbW6RIiLl0JmqVBsuTk4Ma90KZ6fT/rOMHmWbjzU/A3Z/bV5xIiIVoFCVaimvqIjnlq9gfdIRaH+jrfH3RXAy1dzCRETKoVCVaumd3zbyxbbtPLH0J/IbtoeA1lBcCDu+MLs0EZGzUqhKtXRL505cEtaYx67ohdXVFWLG2BbEL4eMBFNrExE5G4WqVEserq68fs3V9GwaYWto0BIaX2KbGi7uA00RJyLVkkJVaoT03DyMdjeAkzMc2QKpu8wuSUTkDApVqfa+3L6dIe9/wI8p2RDR29a4W/Ouikj1o1CVai815yTZBQW8vGo1J5sNBCy2MYF1bVVEqhmFqlR74zp3orGvL8dycnhz12Fo3M22YM835hYmIvI3ClWp9txdXHjw8ssA+HjLFo436W9bcPAXPbcqItVKpUI1MTGRQ4cO2d+vX7+e+++/nzfeeMNhhYmcrlfTCDo0akR+cTGzfz8ODduAUQx7F5ldmoiIXaVC9YYbbmDZsmUApKSkcOWVV7J+/XoeffRRnnnmGYcWKAJgsVi4r/ulACzYsZMjYX1tC/YvhYJsEysTEflLpUJ1+/btdOtmu671+eef07ZtW9asWcPHH3/Mu+++68j6ROw6hoRwWXg4xYbBtN9zwC8civJg349mlyYiAlQyVAsLC7FarQAsXbqUYcOGAdCqVSuSk5MdV53I39wTaztb/WHfPpJCrrA1/v49FBeYV5SIyCmVCtU2bdowZ84cVq1axZIlSxg4cCAASUlJNGjQwKEFipwuqmEAA1tGAvDKQQM8A2wz2MSvMLkyEZFKhuqLL77I66+/zhVXXMHo0aOJiYkBYOHChfZu4Zps5syZREdH07VrV7NLkTLc0bULAD8fOMiJJv1sjXu+gZISE6sSEQGLYVRuENXi4mIyMzPx9/e3t8XHx+Pp6UlgYKDDCjRTZmYmfn5+ZGRk4Ovra3Y5cpoJC79h9cEEbm7XionpH9luVur+gG18YBERB6toHlTqTDU3N5f8/Hx7oB48eJDp06ezZ8+eWhOoUr3d1KEDAMsSkihpdupsVTcsiYjJKhWqV199Ne+//z4A6enpXHLJJbzyyitcc801zJ4926EFipTlkrDGTBk4gC9vGI1T8yvBYoGj2yAzyezSRKQOq1Sobtq0icsvvxyAL7/8kqCgIA4ePMj777/P//73P4cWKFIWi8VC/8gWWF1cwKshNOpkW/CHzlZFxDyVCtWTJ0/i4+MDwI8//siIESNwcnLi0ksv5eDBgw4tUORciktKSA/tZXsTv9z27KqIiAkqFaotWrTgq6++IjExkcWLF9O/v20s1qNHj+qGHqlS248c4ZoPP+bu9QkYXkFQeBIS1phdlojUUZUK1SeffJIHH3yQiIgIunXrRmxsLGA7a+3YsaNDCxQpT4iPD0eyszmYkUl24562xv1LzS1KROqsSj9Sk5KSQnJyMjExMTg52bJ5/fr1+Pr60qpVK4cWaRY9UlMzrE88RNugQDxLcuGb8baB9vu/BPXCzS5NRGqJi/pIDUBwcDAdO3YkKSmJw4cPA9CtW7daE6hSc3QLa4ynmxu4+0GobWAIzbUqImaoVKiWlJTwzDPP4OfnR3h4OE2aNKFevXo8++yzlGhUGzGJYRgkh56avSbhF8g9YW5BIlLnuFTmQ4899hhvv/02L7zwAj169MAwDFavXs3TTz9NXl4ezz33nKPrFClXak4Od379DUmZmaxs3gLnE/vg0DqIHGh2aSJSh1QqVN977z3eeust++w0ADExMYSGhnLXXXcpVKXKNfD0pNgoIaewkE3O4XRlHySsVqiKSJWqVPdvWlpamddOW7VqRVpa2gUXJXK+LBYLY05N7DDjkAXD4gzH90DqbpMrE5G6pFKhGhMTw4wZM85onzFjBu3bt7/gokQqY3BUS7zd3NiSUciRhn/esPStuUWJSJ1Sqe7fKVOmcNVVV7F06VJiY2OxWCysWbOGxMREFi1a5OgaRSrEw9WVoa2i+GTrNj7ICuYhgORNthls3LzNLk9E6oBKnan26tWLvXv3Mnz4cNLT00lLS2PEiBHs2LGDuXPnOrpGkQob2bYNAJ8ezKDQpzGUFEHiWpOrEpG6otKDP5Rly5YtdOrUieLiYkdt0lQa/KFmum3eAjYmJTGrFXQ/sQICoqDPs2aXJSI12EUf/EGkurq2XVsAXj3kjIEFUvdAuiZ6EJGLT6EqtU6f5s3w9/BgT3YxR/3b2Rp3zjO3KBGpExSqUuu4OTszPLo1AG9kN7E1Hl4PuenmFSUidcJ53f07YsSIcpenp6dfSC0iDjOyTRvmbtzEvENZPNyyGdaM/ZCwCqKGml2aiNRi5xWqfn5+51x+8803X1BBIo4Q6udL9/AmrD6YwCpLBP3YDweWQ8shYLGYXZ6I1FLnFap6XEZqkhs7dKBlgwBaR0XAilWQmQgn/oD6LcwuTURqqUoN/iBSE8Q2CSO2SZjtTegltplr9v0I3RSqInJx6EYlqRv+HFg/YTXkZZpbi4jUWgpVqfXWJiRy/5rfyfMJh5JC2L/U7JJEpJZSqEqtN3/HDpbHx7OYSFvDHz/ahi8UEXEwXVOVWm90+/Y09PKiXXRLWL0BctNsz62GdTe7NBGpZWr9meq3335LVFQUkZGRvPXWW2aXIyboFBrCwz0vp1lAEDS/0ta47VMoLjC3MBGpdWp1qBYVFTFx4kR+/vlnNm3axIsvvqhJ1Ou6qCHg4Q/ZKXBgmdnViEgtU6tDdf369bRp04bQ0FB8fHwYPHgwixcvNrssMcm2lCM8uGQlCcFX2Br2LQbHTdIkIlK9Q3XlypUMHTqUkJAQLBYLX3311RnrzJo1i6ZNm+Lu7k7nzp1ZtWqVfVlSUhKhoaH2940bN+bw4cNVUbpUQwt37WLpH3/wvyQruLhD5iHbtVUREQep1qGak5NDTEwMM2bMKHP5Z599xv33389jjz3G5s2bufzyyxk0aBAJCQkAlDVVrEVD1NVZN3bsgAVYmpBMWuPetsZtn0JJ7Zj/V0TMV61DddCgQUyePPmsA/lPnTqV2267jdtvv53WrVszffp0wsLCmD17NgChoaGlzkwPHTpEo0aNzvp9+fn5ZGZmlnpJ7RFerx59mzcHYHZGMLj5QNZhiF9hcmUiUltU61AtT0FBARs3bqR///6l2vv378+aNWsA6NatG9u3b+fw4cNkZWWxaNEiBgwYcNZtPv/88/j5+dlfYWFhF3UfpOqN7dQBgAV748lqfpWtccfnUKQ7gUXkwtXYUE1NTaW4uJigoKBS7UFBQaSkpADg4uLCK6+8Qu/evenYsSMPPfQQDRo0OOs2H3nkETIyMuyvxMTEi7oPUvXaBQfTKaQRRSUlzM0IAM8GtudW9/1gdmkiUgvU2FD909+vkRqGUapt2LBh7N27l3379nHHHXeUuy2r1Yqvr2+pl9Q+4zp1AuDzHXvIjTp1aWH3AijIMbEqEakNamyoBgQE4OzsbD8r/dPRo0fPOHsVOd1lEeE08/cnu6CAzzP9wbexLVB3f2V2aSJSw9XYUHVzc6Nz584sWbKkVPuSJUvo3l3Dz8nZOVks3Hzq2up7m7eS2/o624LfF8HJVPMKE5Ear1qHanZ2NnFxccTFxQFw4MAB4uLi7I/MTJw4kbfeeot33nmHXbt28e9//5uEhATGjx9/Qd87c+ZMoqOj6dq164XuglRTV0VFEV6vHmm5ucxOKIKA1lBcCNs/N7s0EanBLEZZD3NWE8uXL6d3795ntI8dO5Z3330XsA3+MGXKFJKTk2nbti3Tpk2jZ8+eDvn+zMxM/Pz8yMjI0PXVWmhVfDz3fPMdLk5OfDP4EhpteAEsFuj3IvhHmF2eiFQjFc2Dah2qZlOo1n4TFn7D6oMJ9GoawasN9kLiWqjXFPo9B06axElEbCqaB9W6+1fkYnvo8su4NCyMCZdeAh1uATcvSD8Au74yuzQRqYEUqlKnRfj7M+eaYbQMCACPetDxVtuCXfMg/aCptYlIzaNQFTlNTnA3COliGw942ydmlyMiNYxCtQy6+7fuySsqYsrKVQx5/0PSW44CLJC8CVJ3m12aiNQgCtUyTJgwgZ07d7JhwwazS5Eq4urkxOakZE7k5fHjkZMQdqltwS8vQkG2ucWJSI2h2xtFAGcnJ57q25uj2Tn0bBoBhS0gPcE2i83eRdD2OrNLFJEaQGeqIqe0atjQFqgAru4QfWpc4F3z4JAmMxeRc1OoipThSHY2bx1xxwjvBYYBWz6AkiKzyxKRak6hKvI3uYWF3PDZ58xYt55F3peD1RdyjsCmd2wBKyJyFgrVMuju37rNw9WVMTExADz/yzpOtL4BsMD+pbD9M3OLE5FqTcMUlkPDFNZdxSUl3DJvAVtTUujWOJQ57b1x2vgGOLnCkNngrv8eROoSDVMocgGcnZx49sq+uLu4sP7QYd474Q/1m0NJIWx8wzY4hIjI3yhURc4ivF49Hry8BwCvrv2V3/x7gsUZDq/XhOYiUiaFqkg5RrZpw+iY9gBMWHeQtLbjbAt2zoOMQ+YVJiLVkkJVpBwWi4UHL+tB59AQ8ouLeTHegEadbI/X/PoqFJw0u0QRqUYUqiLn4OzkxMOXX44FWLzvD7aGDgOrH2QchNUvQbGeXxURG4WqSAVENQzgmuhoAO7/+VdSOt0LLh5wbIftURsRERSqZdJzqlKWBy7vQauGAaTl5vLPZVsoaHNqPODtn8LxveYWJyLVgkK1DJqlRsri7ebGzGFDaeTjQ5fQUPLCeoJPIyg8aesGLsw1u0QRMZkGfyiHBn+QsuQWFuLh6mp7U5ANS/4LOUeheX/ofLu5xYnIRaHBH0QuEnugAgXOHiRFXm9788ePsP9nk6oSkepAoSpSSZn5+dz59UJGr9jDiaZX2Rp/mwNbPzK3MBExjUJVpJJcnZwoKCqmqLiEfcF9oFk/24LdX8P2z8EoMbdAEalyLmYXIFJTebi68urQq0jNyaFlQACE3WEbGzh+Bez80jZlXORAs8sUkSqkM1WRC1Dfw8MWqKesbTiIwsbdbW+2fgCHdQe5SF2iUBVxkMOZmdz//Y+MSwiiuH4kFBfaHrVJ2mh2aSJSRRSqZdDgD1IZGXl5eLq6suNYGrentafIO8S24JcXIWEN6Ok1kVpPz6mWQ8+pyvmKS07mnm++Iys/nyv8CpnqswWnwpy/VrjsPxDS2bwCRaRS9JyqiAk6NGrER9eNIszPj+UZrozM6k2eb9O/Vti90LziROSiU6iKOFiTevWYMXQIgV5eHMjI4sbkFqTXb2tbmLoL4t7X4zYitZRCVeQiCPevx2ejr6dJPT/2ZRdwxa4GbLM0si3c+60tWIsKzC1SRBxOoSpykfh7ePDuqJEMiYoC4NaUFkzPibQt/H0RzL8R0g+aWKGIOJpCVeQiqu/hwbNX9mXuyOG0Dg7h3ewQPs2P+GuFHx+CHV+aVp+IOJZCVeQis1gsdAwJYdbVQ4kObMgrmRFsj54AzqcG5t/xOez6CvKzTK1TRC6cHqkphx6pEUfLKypi97FjdGjUCE4ep2TdTJyObbct9KgPV74I7n7mFikiZ9AjNSLVkLuLiy1QATwb8IHvYCZnRnLS4g65afDLC5B52NwiRaTSFKoiJjqem8+83BDWNb8FXL0g7Q/48UFIXAvpCRqFSaSGUfdvGWbOnMnMmTMpLi5m79696v6Vi+qPtDSa+vvjlHMUNr4BR7bZlxlt/4HFvR4EtgXvQPOKFKnjKtr9q1Ath66pSlXLzc9l2ZePM9g18cyFnW6H5leCxVL1hYnUcbqmKlIDObu4kdF2LE+6j2J7oU/phZvegpTN5hQmIhWiM9Vy6ExVzLQv5TDLl83lduetpdpPthiGZ4s+4BtiUmUidY/OVEVquBbBofxj5MO81fBm4kr+mgjdc99C+OF+lq9ZYGJ1IlIWnamWQ2eqUp0c2votjXe/X6pthUtrOl55H74+9U2qSqRu0I1KDqBQleqmqKiQ3TtXYd31GZGWE38t8AmFyEHQtPdfIzWJiMMoVB1AoSrVVV5REet/+ZhuGWtwz0+zt6eWuHG03R1ER18OxYXg4mZilSK1h0LVARSqUiMc2QorJtvfFhoWXLwCsJw8Bn5hGG1HYwntYmKBIjWfblQSqSuC2sO1n1EYGAOAq8WwBSpARiKW1VP43+o1ZOblmVikSN3gYnYBIuIAFguuPR+BzEOwYSa4eMKxHfbFA+LfZMjmGC4L8iG2oQ8RrXrSMiAAq4v+BIg4krp/y6HuX6mxigpg7VRI3mRvii/yIMIlF4AZ2RGsdo3m2cFDadGggVlVitQYuqbqAApVqfHys+CPJbDnayjMLbUorsCXZYVBeDVoinNIR4ZHt6a+p6dJhYpUbwpVB1CoSq2RkQBrpkJWUpmL7znRlrVFATzZpw/DWreq4uJEqr+K5oEuqIjUBX5NYMArkJ8BuxfC8b0Y6QexlBQC8Jr/dg4b3jSo39v+kRLDwEmD94ucF52plkFTv0mdUJAN2z+DfYtPa7RA2KWcaHkt/1j4E1e2aM6/e3TH2cn2oEBWfj7uLi64OjubU7OISdT96wDq/pVar7gA/lgKSRvh6F/zuJZgYcHJYFZ7dmbqdTcDsPFwErfPX0C74GBev2YYAB6uGr1J6gaFqgMoVKVOMQzb1HIb34Y/n3MF2/yt0aPIa3kNl855w97s7+7Ou6NGEu5fr+prFaliClUHUKhKnVRcAClbYPfXcHzvX+0N2/B7iS/T/8hhdb4/AD5WK2Ni2nNDhxh8rVaTCha5+BSqDqBQlTrNMGDfD7DlQzh1QxOAgYVdjYfw3J5sBrKXebkhFHoFMziqJQXFxXi7uTGuU0ddd5VaRaHqAApVkVMyDsH+JbD/ZyjOL7UortCPcWkdSrUNjIxkcv9+uDhpJFSpHTT2r4g4jl9j6HgLjHgPWg4ptaiDawa/hvzGrNBk5jX4jTDnXLBgD9QSw2DPsVRO//d7iWGw6XAShcXFVbobIhebzlTLoTNVkbMoKoDfF8G2j8tcnOkRjG/ncRDSid3HjvGPTz8nzM+Pr28ag5PFwtRfVvP+5jgGR7Xk//pfWbW1i1SCzlRF5OJxcYPW18BVs6DtP6BeRKnFvrkp8MsLsPEtEtOO4e7iQlN/f5wsFgzDsA/kv2jPXhb//nvV1y9ykehMtRw6UxWpIMOwPYaz7VNI3VP6kRzPBpS416fQxQtrm2ugQStyCgv41zvT2FXkg8XJmc6hIYzv1pWOISGm7YJIeXSjkgMoVEUqKWkTbHkPspLLXu7fFE4cYJFTWx5N/muWnOjAQO7vHkuXxqEaIlGqFYWqAyhURS5QUZ4tYPcvhaPby1xl/aUvM23Nr+w69tfZbeuGDRnZtg0DW0bi7eZWVdWKnJVC1QEUqiIOlHMUEtfC1o9Ktwe0osSvCUkWf7bH72R+qivr830AcLJYeLjn5fyjfTsTChb5i2apEZHqxSsQWl0NEVfA2mm2ruG8dEjdjVPqbhoDjYH+QfX4oPEdfLljF0mZWcSGhdk38fr6DThZLIyJaY/nqTPYlKwsikpKaOznZ8ZeiZSiUBWRquXuB72ftv2ckQjJm2yvE/vBMHDKT2ds/KuMHfYcJ04k4e/rbf9oXHIyaxMSaejlxTXRrVm0Zy+P/rgEJ4uFBy/vwbVt22okJzGVun/Loe5fkSpkGLB5rm1oxNM17w+dbwfgu917eHfTZt4eORxfq5WikhJun/8Vccm2G6LC/Px4qm9v2gYF4e6icwZxHF1TdQCFqkgVK8y1hWr8CshK+qvd3Q88A6FZHwjtBlYf+6L8oiLu/24RaxMS7W0NvTy585JLGBzVUuEqDqFQdQCFqohJigvhwDJbwGYeKr3MzQt6Pm67RpudDD6Nwc2TRXv28tKqX8gpKKDg1PCHvlYrjf386BHehD7NmtGqYQAWPaojlaBQdQCFqojJDANOptqmotv7zdmfe/UOgq4ToH4LjuTmMW31Gpb9sZ/8v40t/O8e3RnbqSMAR7KzOZGbS1SAglbOTXf/ikjNZ7GAV0No3s/2KjgJK56FE3+UXi/7CCx7ErAQ1PxKXuj9D7Ku6MX2I0fZd/w4S/b9wd7UVPzc/5rzdem+P3hp1S8MiYpicv9+VbtfUmspVMswc+ZMZs6cSbFm0BCpXtw84YonIXkzeAfDwZW2gf3tDPjjR0iJw6fLv4ht0o7YJmHc1LED+UVFlJzWMedzalL1NkGB9rbNScl4uroS1TCgqvZIahl1/5ZD3b8iNUB+JhQXwfoZZY/a5BVou7mp8KQtiFtfY1/0e+pxmtb3t09T9+/vFrFs/wHaBgUyIDKSUF9fGvn46FqsqPtXROoI66k/cD0egsPrIKAVrJ1ue+4VbCM57f32r/WdXGzh6h9BpI87nDaRuu+ps9ftR46y/chRe3vLgAbc1qUz3Ro3xtdqxVmTr8tZ6Ey1HDpTFamhCrIhfiUc22m7k/hkKmQmlr1utwkQ0cv+NiE9ne/3/k5ccjI5BQX8fvw4uYVF9uVOFgt9mzfj7ksvJdjH2z6NndRuuvvXARSqIrVIXiZsfhty0yH9gG2w/z81bG07003aBB71IKi9fVFGXh6fbNnKvB07OJZzstQmm9evzxc3/KNCM+oUFBfjptGeaiyFqgMoVEVqqZStsHLy2ZeHdLGNUxwQ9VdbSQmFWcn8mJLNk0t+otgwmHP1MC5tYhub+MvtO9iTmsrtXToT5O1danMvrfqFj+K2cEnjxrw0aAC+7u4XY6/kIlKoOoBCVaQWy8uE/AzbIBPxy6Ag58x13LygxSDwDbVNvr7vB7j0PjIDO+Pp5oazxWK/gel/a9byzsZNPHpFL65r1xaAP9LS+PmP/cz8dZ19k72aRvDqkKuqZBfFcXSjkohIedx9ba8ON0P7MbZnXYtyYf1M20hNJcW2oN35ZenP/foqvr0ehwatwOWvuV4vDQtjw6HDND7tD+7/1qxlxYF4AC6PCGdtQiIrDsSz48jRUo/ySO2hM9Vy6ExVpI4qKoDf5tiC1upjm0Xn71w9oPVIaNQB/JqUuZmnlv7EmoRERrVtwx1du/D4kqV8t2cvIb4+PNOvL11CQ8stIyPPdt3XT93FplP3rwMoVEUEgPQE2PSmrQu4LC2HwJEttmdiuz9U6jEdwHY3sos7h9PTuX3hIpKzsgBo1TCAa1q3pk/z5gR6ewFgGIa9S3nZ/v08v3wlH153rX25mEOh6gAKVREpJTfdNrG6dyD8MBFy085cJ7ynbVadehG2ASf2/fjX4zxWX3J6PcvUTbtZuGs3hSUlAMQ2CWP21cM4kHaCsV/OY8awIbQPDmbG2l9567eNxDQKZvawoXi6uZGUmUmJYWhS9iqma6oiIo7mUc/2Auj3POSdgLj3bc/D/ungyrN/Pj8Tr70LeKLPXdwTG8v3e/fy9a5deLnZrs2+vXEjmfn57DueRvvgYAZHtWTuxk1sSU6h++tv0imkEXHJKbg5O/PZ6OsJr1fvYu2pVJLOVMuhM1UROafcdNtIToFtIDPJFqqH1599fSdn6P6gbQae0C6lFmXl5/PMz8uICgjg9q62ZbN+XccbG347YzNhfn70j2zB7V064+Hq6sg9kjKo+9cBFKoiUiklRfD79xC/HDJOdf2GdIETByD3+F/rRQ2DsO7g29g2McDeb21zxRpFthGhCk9CYFvi68Vw+/yvCPbx5oaYGJ5a+pO96xjg3VEj6NCoUdXuYx2jUHUAhaqIXLDUPbbxiX0aQeJaWDut/PUDos68IWrUpxQDzrnH4dgu9ni24vmVvxCXnEx0YEM+uu5aLBYLOQUFzPx1HeMv6WYfx1gcQ9dURUSqg9NHZQqLhZPHbeHqFwZHttrGJT5dWXcYZyTg7B9he8znyDaiWo/g3VH/4IPNcaTm5HAyeRtewdE89P1i1iQk0LVxKL2bNQMgt7AQi8XCtpQU2gYFqav4IlOoiohUpaghttefMhJhy4eQsvnsn1nyMHgG/BXAu+YDcJNxEhr4wy+vQuQg7okdQnz6CX5PPW4P1RdXruKrnbsACPTy4oHLezAgMhKAEsOo0LjFUnHq/i2Hun9FpEoYhu0M1SvQ9phOQZZtbtg935zfdpr3pyigNVmBnfD38ADg1dVrmbup9OAVnUND8HGzsv7QIRr7+nJTxw4MaRWlOWPLoWuqDqBQFRFT5aXDrq8guANseQ8yD0O9prZZdQ6usoVvWQZOBVdvsFgodPVmy+5faXP4OxZZu/DcthRKyviz39jXl2vbteXatm3wdHMrY6N1m0LVARSqIlJt5KbD7q+geX/wDbG1bZgDB34++2csTtDlX7Bznm2ydlcPEnq/yq+Jh0jJysLbzY28oiLe3PAbfwbB56Ovp2VAAAApWVlYLJYzZt2pi3SjkohIbeJRDzqOK93WfoxtAIrAdraz1xP7YftnkJ9pW26UwIbZf61fmEuT9C00adIcXINt88cGRnNFs6Ys3fcHK+PjCfCyDYdYVFLC/d8t4lBGJktuHYeHqytbklNYceAAEf7+WJ2dOZ57kiFRUZrK7jQKVRGRmsrqA5c/8tf7+s2h8SW2IRSL8sFiKT0ZO9hm4Tmdhz/RPR4muoUv9156nW1wCiAhPZ2ikhJyCgpYfuAAg1q25Eh2Nu9sLH199oPNW+jeJIx9aWm4O7vwwsD+9uu5dZG6f8uh7l8RqZFy020DSBgGHFoHWUm29sTVUJhb/mdbXW3rKm58KYTFUlyYh3PyRgjpSomzK/d+8x3HT55k97FjnC08Gvl4M7ZjR/4R096Re2UqXVN1AIWqiNQqRQWAAVveh8O/QUQv2PM14ARG8Znr1wuH9IO2n8O6234O7Qrtb4D0g+zKLuKbXbtZnnQcd2cX9p/4c4IBC/U9PPj59lvtm/q/5SvwcnXjvh6xF3svLwpdUxURkdL+nFS98z9tL4DoUbZu4pQ4iHvPNsPO8d9ty/4MVIDENbb/3X0YkjdDxkFaA62dnHm4378hqAM53z9AES784BlLSUai7UzZYuFIdjZfbNuOAbQLDqJP82YYhsHyAwcoKTEoKikhNrxJrRgFSmeq5dCZqojUSSdTYe8iKM63TRBQlF+57bS/EdzrkVEviiUHk+lyYhVNCg9j6fEw0zbu5P24LfZVfaxW7rykK04WJ9oHB7E5KZll+w9QXFLCPbGX0ik0xEE7Vznq/nUAhaqICLYz2IOroM11tvd7v4HsI+e3DScX20QDp6xxa8vjqY252jeDa0u28NCJKHYUlf139vtxN9PIxwcoPYl7VVKoOoBCVUTkLNIPws4vod1oKMyzXZs9EQ/ZyRXfRsshtpl5gFSXBtxf1IcSw2DP0aN4WorwsJQQ6JTPDbF9GNy+K5l5OSz9bjpN2g6gS+tukByHsfsrUlrfREDDCFwxYN3/wK8JtBnl0N1VqDqAQlVE5Dzt/wl+ewNaXgUNIqFBFCRvgl3zbJMJlMcn5K87lf/OOxiyUwA4bvGmwfA5MP9GAFbn+/MSPbk/3JkrUmzjIv/L+Tr+07sPzerXd8huKVRPM3z4cJYvX07fvn358ssvK/w5haqISCUYJbbRnEq1GVBcCKtftHUDu/vbJhPIz4T8jAv+yrQSV1wpwcfJdhfzviJPXslqzsg+o+kX2eKCt69QPc2yZcvIzs7mvffeU6iKiFQ3BSfh+F7bWeqJ/bbrr6cPv9jrCTiwDBJ+Oe9NZ5U44+YThLXdKGhyWaVL1CM1p+nduzfLly83uwwRESmLmyc06mB7/ck7CLZ9YhsxKqgdBLaxPTf7xxLb4BQVkIcLPk5FkJNU6iapi8n0UF25ciUvvfQSGzduJDk5mQULFnDNNdeUWmfWrFm89NJLJCcn06ZNG6ZPn87ll19uTsEiInLxtboGWgwE11NDHlqcbKM9tbzKdm3WswHErwDPhrbHfuqF27qXs5LA2Q2a9cPdtzGcPGpb3zesSso2PVRzcnKIiYnhlltuYeTIkWcs/+yzz7j//vuZNWsWPXr04PXXX2fQoEHs3LmTJk2aANC5c2fy8898jurHH38kJKTizzbl5+eX2k5mZmYl9khERC6YxfJXoJ7OycV2FgvQrK/tf4PLGQ7Rt7HtVUVMD9VBgwYxaNCgsy6fOnUqt912G7fffjsA06dPZ/HixcyePZvnn38egI0bNzqklueff55JkyY5ZFsiIlL3OJ17FfMUFBSwceNG+vfvX6q9f//+rFmzxuHf98gjj5CRkWF/JSYmOvw7RESk9jL9TLU8qampFBcXExQUVKo9KCiIlJSUCm9nwIABbNq0iZycHBo3bsyCBQvo2rXrGetZrVastWDsSRERMUe1DtU//X1IqvMdpmrx4sWOLklEROQM1br7NyAgAGdn5zPOSo8ePXrG2auIiIjZqnWourm50blzZ5YsWVKqfcmSJXTv3v2ife/MmTOJjo4us4tYRETkbEzv/s3Ozmbfvn329wcOHCAuLo769evTpEkTJk6cyE033USXLl2IjY3ljTfeICEhgfHjx1+0miZMmMCECRPsI2iIiIhUhOmh+ttvv9G7d2/7+4kTJwIwduxY3n33Xa6//nqOHz/OM888Q3JyMm3btmXRokWEh4ebVbKIiEiZ6sTYv5WlsX9FRAQqngfV+pqqiIhITaJQFRERcRDTr6lWRzNnzmTmzJkUFdlmNdAYwCIiddufOXCuK6a6plqOQ4cOERZWNTMbiIhI9ZeYmEjjxmcfoF+hWo6SkhKSkpLw8fE5rxGc/i4zM5OwsDASExN1w9M56FhVnI5Vxeg4VZyO1dkZhkFWVhYhISE4OZ39yqm6f8vh5ORU7r9Izpevr6/+Q60gHauK07GqGB2nitOxKltFxi3QjUoiIiIOolAVERFxEIVqFbBarTz11FOaVq4CdKwqTseqYnScKk7H6sLpRiUREREH0ZmqiIiIgyhURUREHEShKiIi4iAKVREREQdRqFaBWbNm0bRpU9zd3encuTOrVq0yu6QqtXLlSoYOHUpISAgWi4Wvvvqq1HLDMHj66acJCQnBw8ODK664gh07dpRaJz8/n3vuuYeAgAC8vLwYNmwYhw4dqsK9uPief/55unbtio+PD4GBgVxzzTXs2bOn1Do6VjazZ8+mffv29kEKYmNj+f777+3LdZzO7vnnn8disXD//ffb23S8HMiQi+rTTz81XF1djTfffNPYuXOncd999xleXl7GwYMHzS6tyixatMh47LHHjHnz5hmAsWDBglLLX3jhBcPHx8eYN2+esW3bNuP66683GjVqZGRmZtrXGT9+vBEaGmosWbLE2LRpk9G7d28jJibGKCoqquK9uXgGDBhgzJ0719i+fbsRFxdnXHXVVUaTJk2M7Oxs+zo6VjYLFy40vvvuO2PPnj3Gnj17jEcffdRwdXU1tm/fbhiGjtPZrF+/3oiIiDDat29v3HffffZ2HS/HUaheZN26dTPGjx9fqq1Vq1bGf//7X5MqMtffQ7WkpMQIDg42XnjhBXtbXl6e4efnZ8yZM8cwDMNIT083XF1djU8//dS+zuHDhw0nJyfjhx9+qLLaq9rRo0cNwFixYoVhGDpW5+Lv72+89dZbOk5nkZWVZURGRhpLliwxevXqZQ9VHS/HUvfvRVRQUMDGjRvp379/qfb+/fuzZs0ak6qqXg4cOEBKSkqpY2S1WunVq5f9GG3cuJHCwsJS64SEhNC2bdtafRwzMjIAqF+/PqBjdTbFxcV8+umn5OTkEBsbq+N0FhMmTOCqq66iX79+pdp1vBxLA+pfRKmpqRQXFxMUFFSqPSgoiJSUFJOqql7+PA5lHaODBw/a13Fzc8Pf3/+MdWrrcTQMg4kTJ3LZZZfRtm1bQMfq77Zt20ZsbCx5eXl4e3uzYMECoqOj7X/kdZz+8umnn7Jp0yY2bNhwxjL9d+VYCtUq8Pdp4wzDuKCp5Gqjyhyj2nwc7777brZu3covv/xyxjIdK5uoqCji4uJIT09n3rx5jB07lhUrVtiX6zjZJCYmct999/Hjjz/i7u5+1vV0vBxD3b8XUUBAAM7Ozmf8S+7o0aNn/KuwrgoODgYo9xgFBwdTUFDAiRMnzrpObXLPPfewcOFCli1bVmrqQR2r0tzc3GjRogVdunTh+eefJyYmhldffVXH6W82btzI0aNH6dy5My4uLri4uLBixQr+97//4eLiYt9fHS/HUKheRG5ubnTu3JklS5aUal+yZAndu3c3qarqpWnTpgQHB5c6RgUFBaxYscJ+jDp37oyrq2updZKTk9m+fXutOo6GYXD33Xczf/58fv75Z5o2bVpquY5V+QzDID8/X8fpb/r27cu2bduIi4uzv7p06cKYMWOIi4ujWbNmOl6OZM79UXXHn4/UvP3228bOnTuN+++/3/Dy8jLi4+PNLq3KZGVlGZs3bzY2b95sAMbUqVONzZs32x8reuGFFww/Pz9j/vz5xrZt24zRo0eXeTt/48aNjaVLlxqbNm0y+vTpU+tu57/zzjsNPz8/Y/ny5UZycrL9dfLkSfs6OlY2jzzyiLFy5UrjwIEDxtatW41HH33UcHJyMn788UfDMHSczuX0u38NQ8fLkRSqVWDmzJlGeHi44ebmZnTq1Mn+iERdsWzZMgM44zV27FjDMGy39D/11FNGcHCwYbVajZ49exrbtm0rtY3c3Fzj7rvvNurXr294eHgYQ4YMMRISEkzYm4unrGMEGHPnzrWvo2Nlc+utt9r/P9WwYUOjb9++9kA1DB2nc/l7qOp4OY6mfhMREXEQXVMVERFxEIWqiIiIgyhURUREHEShKiIi4iAKVREREQdRqIqIiDiIQlVERMRBFKoiIiIOolAVkYvCYrHw1VdfmV2GSJVSqIrUQuPGjcNisZzxGjhwoNmlidRqmk9VpJYaOHAgc+fOLdVmtVpNqkakbtCZqkgtZbVaCQ4OLvXy9/cHbF2zs2fPZtCgQXh4eNC0aVO++OKLUp/ftm0bffr0wcPDgwYNGnDHHXeQnZ1dap133nmHNm3aYLVaadSoEXfffXep5ampqQwfPhxPT08iIyNZuHDhxd1pEZMpVEXqqCeeeIKRI0eyZcsWbrzxRkaPHs2uXbsAOHnyJAMHDsTf358NGzbwxRdfsHTp0lKhOXv2bCZMmMAdd9zBtm3bWLhwIS1atCj1HZMmTeK6665j69atDB48mDFjxpCWllal+ylSpcyeJkdEHG/s2LGGs7Oz4eXlVer1zDPPGIZhm2Zu/PjxpT5zySWXGHfeeadhGIbxxhtvGP7+/kZ2drZ9+XfffWc4OTkZKSkphmEYRkhIiPHYY4+dtQbAePzxx+3vs7OzDYvFYnz//fcO20+R6kbXVEVqqd69ezN79uxSbfXr17f/HBsbW2pZbGwscXFxAOzatYuYmBi8vLzsy3v06EFJSQl79uzBYrGQlJRE3759y62hffv29p+9vLzw8fHh6NGjld0lkWpPoSpSS3l5eZ3RHXsuFosFAMMw7D+XtY6Hh0eFtufq6nrGZ0tKSs6rJpGaRNdUReqoX3/99Yz3rVq1AiA6Opq4uDhycnLsy1evXo2TkxMtW7bEx8eHiIgIfvrppyqtWaS605mqSC2Vn59PSkpKqTYXFxcCAgIA+OKLL+jSpQuXXXYZH330EevXr+ftt98GYMyYMTz11FOMHTuWp59+mmPHjnHPPfdw0003ERQUBMDTTz/N+PHjCQwMZNCgQWRlZbF69Wruueeeqt1RkWpEoSpSS/3www80atSoVFtUVBS7d+8GbHfmfvrpp9x1110EBwfz0UcfER0dDYCnpyeLFy/mvvvuo2vXrnh6ejJy5EimTp1q39bYsWPJy8tj2rRpPPjggwQEBDBq1Kiq20GRashiGIZhdhEiUrUsFgsLFizgmmuuMbsUkVpF11RFREQcRKEqIiLiILqmKlIH6aqPyMWhM1UREREHUaiKiIg4iEJVRETEQRSqIiIiDqJQFRERcRCFqoiIiIMoVEVERBxEoSoiIuIg/w9LfdvGDsXD5QAAAABJRU5ErkJggg==", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "from mlcolvar.utils.plot import plot_metrics\n", "\n", "ax = plot_metrics(metrics.metrics, \n", " keys=['train_loss_epoch','valid_loss'],\n", " linestyles=['-.','-'], colors=['fessa1','fessa5'],\n", " yscale='log')" ] }, { "attachments": {}, "cell_type": "markdown", "metadata": {}, "source": [ "### Analyze the results" ] }, { "attachments": {}, "cell_type": "markdown", "metadata": {}, "source": [ "Once the model is optimized, we can directly use it to predict the cv for new data. And of course, we can compare the predicted vs reference data." ] }, { "cell_type": "code", "execution_count": 130, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "Text(0, 0.5, 'Prediction')" ] }, "execution_count": 130, "metadata": {}, "output_type": "execute_result" }, { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "from torch import no_grad # allows for inference without gradients\n", "\n", "# predict data\n", "model.eval()\n", "with no_grad():\n", " y_pred = model(Tensor(X))\n", "\n", "# compare with reference\n", "import matplotlib.pyplot as plt\n", "_,ax = plt.subplots(dpi=100)\n", "ax.scatter(y,y_pred,s=1,color='fessa0',alpha=0.8)\n", "ax.plot(y,y,linestyle='--',color='grey')\n", "#ax.set_xlim(-0.1,1.1)\n", "#ax.set_ylim(-0.1,1.1)\n", "ax.set_xlabel('Reference')\n", "ax.set_ylabel('Prediction')\n" ] }, { "attachments": {}, "cell_type": "markdown", "metadata": {}, "source": [ "Since this is a two-dimensional model potential we can also compare the results in the (x,y) space. To do that we need to compute the isolines of the CV and compare with the true potential. The prediction is actually rather accurate, except for the regions where few if any data points are available. " ] }, { "cell_type": "code", "execution_count": 131, "metadata": {}, "outputs": [ { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "fig,axs = plt.subplots(1,3,dpi=100,figsize=(12,4),sharex=True,sharey=True)\n", "for ax,title in zip(axs,['Reference','Training points','Prediction']):\n", " ax.set_title(title)\n", " ax.set_xlabel('x')\n", " ax.set_ylabel('y')\n", "\n", "plot_isolines_2D(muller_brown_potential,levels=12,max_value=24,ax=axs[0])\n", "pp = axs[1].scatter(X[:,0],X[:,1],c=y[:,0],s=1,alpha=1,cmap='fessa')\n", "plt.colorbar(pp,ax=axs[1])\n", "plot_isolines_2D(model,levels=12,max_value=24,ax=axs[2])\n", "plt.tight_layout()" ] }, { "attachments": {}, "cell_type": "markdown", "metadata": {}, "source": [ "### Deploy the model in PLUMED" ] }, { "attachments": {}, "cell_type": "markdown", "metadata": {}, "source": [ "In order to deploy the model to PLUMED we first need to compile it TorchScript (see `torch.jit`). We will trace the model, meaning that all of the instructions that are executed in an example workflow will be compiled in a python-independent format. Once again we can use lightning wrapper to do so, and just call the method `to_torchscript`." ] }, { "cell_type": "raw", "metadata": {}, "source": [ "model.to_torchscript('model.ptc',method=\"trace\")" ] }, { "attachments": {}, "cell_type": "markdown", "metadata": {}, "source": [ "Then an example input file to evaluate and print the CV from PLUMED could be:\n", "\n", "\n", " # ------------- plumed.dat -------------\n", " p: POSITION ATOM=1\n", " cv: PYTORCH_MODEL FILE=model.ptc ARG=p.x,p.y\n", " \n", " PRINT STRIDE=100 ARG=cv.*\n", " # --------------------------------------\n", "\n", "For complete details on how you should use your data-driven CVs in PLUMED you should check the [PLUMED documentation](https://www.plumed.org/doc-v2.9/user-doc/html/_p_y_t_o_r_c_h__m_o_d_e_l.html)!" ] }, { "attachments": {}, "cell_type": "markdown", "metadata": {}, "source": [ "You made it to the end! This first tutorial was designed to give an overview on the process of training data-driven CVs with `mlcolvar`. You can now navigate through the CV-specific tutorials which describe the different approaches which are implemented as well as information on how to customize the CVs and the training. " ] } ], "metadata": { "kernelspec": { "display_name": "pytorch", "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": "3.10.8" }, "orig_nbformat": 4, "vscode": { "interpreter": { "hash": "1cbeac1d7079eaeba64f3210ccac5ee24400128e300a45ae35eee837885b08b3" } } }, "nbformat": 4, "nbformat_minor": 2 }