Merge pull request #108 from pluflou/add-gps

Add GPModel class

Author: pluflou

environment.yml

@@ -3,7 +3,7 @@ name: lume-model
 channels:
   - conda-forge
 dependencies:
-  - python>=3.9
+  - python>=3.10
   - pydantic>2.3
   - numpy
   - pyyaml

examples/probabilistic_models/custom_model_probabilistic.ipynb

@@ -0,0 +1,190 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "id": "235c92cd-cc05-42b8-a516-1185eeac5f0c",
+   "metadata": {},
+   "source": [
+    "# Creating a Custom LUME-model for probabilistic models\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "id": "56725817-2b21-4bea-98b0-151dea959f77",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from torch.distributions.normal import Normal\n",
+    "import torch\n",
+    "from lume_model.models.prob_model_base import ProbModelBaseModel\n",
+    "from lume_model.variables import ScalarVariable, DistributionVariable"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "79c62b18-7dc1-44ca-b578-4dea5cc4a4b4",
+   "metadata": {},
+   "source": [
+    "## Create a model that returns a list of predictions"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "id": "f96d9863-269c-49d8-9671-cc73a783bcbc",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "class ExampleModel(ProbModelBaseModel):\n",
+    "    \"\"\"Model returns a list of predictions for each output\"\"\"\n",
+    "\n",
+    "    def _get_predictions(self, input_dict):\n",
+    "        \"\"\"\n",
+    "        This method implements the required abstract method for this class.\n",
+    "        It takes the input_dict and returns a dict of output names to distributions.\n",
+    "        \"\"\"\n",
+    "        # Just generate random output here for this example\n",
+    "        # but typically this is where you would adjust the input if needed and\n",
+    "        # call your model on the input\n",
+    "        output_dict = {\n",
+    "            \"output1\": torch.rand(5),\n",
+    "            \"output2\": torch.rand(10),\n",
+    "        }\n",
+    "        return self._create_output_dict(output_dict)\n",
+    "\n",
+    "    def _create_output_dict(self, output):\n",
+    "        \"\"\"This method is not required by the abstract class but typically\n",
+    "        needed to create a distribution type output for each output\n",
+    "        name from the list of predictions that the model returns.\n",
+    "        \"\"\"\n",
+    "        output_dict = {}\n",
+    "        for k, v in output.items():\n",
+    "            output_dict[k] = Normal(v.mean(axis=0), torch.sqrt(v.var(axis=0)))\n",
+    "        return output_dict"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "868fff4d-1f46-48e2-8bd0-c9d831df79e6",
+   "metadata": {},
+   "source": [
+    "### Model Instantiation and Execution\n",
+    "Instantiation requires specification of the input and output variables of the model."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 3,
+   "id": "97946e64-062d-47d4-8d0c-d7e02a335a56",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "input_variables = [\n",
+    "    ScalarVariable(name=\"input1\", default_value=0.1),\n",
+    "    ScalarVariable(name=\"input2\", default_value=0.2, value_range=[0.0, 1.0]),\n",
+    "]\n",
+    "output_variables = [\n",
+    "    DistributionVariable(name=\"output1\"),\n",
+    "    DistributionVariable(name=\"output2\"),\n",
+    "]\n",
+    "\n",
+    "m = ExampleModel(input_variables=input_variables, output_variables=output_variables)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 4,
+   "id": "50aae4be-0d6e-456f-83e8-3a84d6d78f84",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "input_dict = {\n",
+    "    \"input1\": 0.3,\n",
+    "    \"input2\": 0.6,\n",
+    "}\n",
+    "out = m.evaluate(input_dict)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 5,
+   "id": "9a74c70a-4d9a-443f-820d-69d111e574ed",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "{'output1': Normal(loc: 0.4858802855014801, scale: 0.3480694591999054),\n",
+       " 'output2': Normal(loc: 0.5287243127822876, scale: 0.28792139887809753)}"
+      ]
+     },
+     "execution_count": 5,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "out"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 6,
+   "id": "301aa223-f53f-498f-8b31-ed1008594f87",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "(tensor(0.4859), tensor(0.1212))"
+      ]
+     },
+     "execution_count": 6,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "out[\"output1\"].mean, out[\"output1\"].variance"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "35ea12da-d0c6-49cc-8a00-2b096fc7248b",
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "63d56fd9-8c25-4371-921b-968773376203",
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3 (ipykernel)",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.10.16"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}