{ "cells": [ { "cell_type": "markdown", "id": "a1ea94bd", "metadata": {}, "source": [ "# My First World Machine" ] }, { "cell_type": "markdown", "id": "8b00b00a", "metadata": {}, "source": [ "[![Open in Colab](https://img.shields.io/badge/Open%20in%20Colab-F9AB00?style=for-the-badge&logo=googlecolab&color=525252)](https://colab.research.google.com/github/H-IAAC/WorldMachine/blob/main/examples/My%20First%20World%20Machine.ipynb) [![Open in Github](https://img.shields.io/badge/Open%20in%20Github-100000?style=for-the-badge&logo=github&logoColor=white)](https://github.com/H-IAAC/WorldMachine/blob/main/examples/My%20First%20World%20Machine.ipynb)" ] }, { "cell_type": "markdown", "id": "665f3441", "metadata": {}, "source": [ "World Machine is a research project that investigates the concept and creation of computational world models.\n", "It is also the name for the project architecture and training protocol.\n", "\n", "In this example, we will create a World Machine and train it on an example dataset.\n", "\n", "Let's start by importing the `world_machine` package and other required modules:" ] }, { "cell_type": "code", "execution_count": 1, "id": "2e361f4d", "metadata": {}, "outputs": [], "source": [ "try:\n", " import world_machine\n", "except:\n", " !python3 -m pip install world_machine" ] }, { "cell_type": "code", "execution_count": null, "id": "9dc20c3b", "metadata": {}, "outputs": [], "source": [ "import matplotlib.pyplot as plt\n", "import seaborn as sns\n", "import torch\n", "\n", "import world_machine as wm" ] }, { "cell_type": "markdown", "id": "292ecb16", "metadata": {}, "source": [ "And by checking if there is an available GPU. This will speed up the example training and evaluation:" ] }, { "cell_type": "code", "execution_count": 3, "id": "81469dfe", "metadata": {}, "outputs": [], "source": [ "device = 'cuda' if torch.cuda.is_available() else 'cpu'" ] }, { "cell_type": "markdown", "id": "d6dc24dc", "metadata": {}, "source": [ "## Creating our Dataset" ] }, { "cell_type": "markdown", "id": "29fdb4ca", "metadata": {}, "source": [ "For this example, we will create a model that learns the behavior of a sine function and predicts its future values. We define the data, creating our time vector, and defining random phases and frequencies:" ] }, { "cell_type": "code", "execution_count": 4, "id": "80a4cb24", "metadata": {}, "outputs": [], "source": [ "t = torch.linspace(0, 99, 100)\n", "phase = torch.lerp(torch.tensor(0.0), torch.tensor(2*torch.pi), torch.rand(1000))\n", "frequency = torch.lerp(torch.tensor(0.1), torch.tensor(1.0), torch.rand(1000))\n", "\n", "x = frequency[:, None] * t[None, :]\n", "x += phase[:, None]\n", "data = torch.sin(x)" ] }, { "cell_type": "markdown", "id": "4fbdf3b7", "metadata": {}, "source": [ "We can plot some dataset samples:\n" ] }, { "cell_type": "code", "execution_count": 5, "id": "0f13a5eb", "metadata": {}, "outputs": [ { "data": { "text/plain": [ "Text(0.5, 0.98, 'Data Samples')" ] }, "execution_count": 5, "metadata": {}, "output_type": "execute_result" }, { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "fig, axs = plt.subplots(3, 3)\n", "\n", "axs_flat = axs.flatten()\n", "\n", "for i in range(9):\n", " axs_flat[i].plot(data[i])\n", "\n", "plt.suptitle(\"Data Samples\")" ] }, { "cell_type": "markdown", "id": "ab2300b1", "metadata": {}, "source": [ "To train a World Machine using our data, we need to create a `WorldMachineDataset`. It has methods for loading our data in the format the model and protocol expect. We are going to make a dataset with only one sensory channel named \"data\" that contains our sine signal:" ] }, { "cell_type": "code", "execution_count": 6, "id": "642a0afc", "metadata": {}, "outputs": [], "source": [ "class FirstDataset(wm.data.WorldMachineDataset):\n", " def __init__(self, data:torch.Tensor):\n", " super().__init__(sensory_channels=[\"data\"], \n", " size=len(data))\n", " \n", " self._data = data\n", "\n", " def get_channel_item(self, channel, index):\n", " if channel == \"data\":\n", " x = self._data[index]\n", " y = torch.roll(self._data[index], -1)\n", " return x[:-1].unsqueeze(1), y[:-1].unsqueeze(1)" ] }, { "cell_type": "code", "execution_count": 7, "id": "be95aef5", "metadata": {}, "outputs": [], "source": [ "train_dataset = FirstDataset(data[:700])\n", "val_dataset = FirstDataset(data[700:])" ] }, { "cell_type": "markdown", "id": "2a535a8c", "metadata": {}, "source": [ "Then, we create dataloaders for our data:" ] }, { "cell_type": "code", "execution_count": 8, "id": "264a7ee8", "metadata": {}, "outputs": [], "source": [ "batch_size = 32\n", "dataloaders = {\"train\":wm.data.WorldMachineDataLoader(train_dataset, batch_size, shuffle=True, drop_last=True),\n", " \"val\":wm.data.WorldMachineDataLoader(val_dataset, batch_size, shuffle=True, drop_last=True)}" ] }, { "cell_type": "markdown", "id": "e2c0a4cd", "metadata": {}, "source": [ "We can inspect one batch. Observe that the data is structured using PyTorch's TensorDict:" ] }, { "cell_type": "code", "execution_count": 9, "id": "abe6cc5c", "metadata": {}, "outputs": [ { "data": { "text/plain": [ "TensorDict(\n", " fields={\n", " index: Tensor(shape=torch.Size([32]), device=cpu, dtype=torch.int64, is_shared=False),\n", " inputs: TensorDict(\n", " fields={\n", " data: Tensor(shape=torch.Size([32, 99, 1]), device=cpu, dtype=torch.float32, is_shared=False)},\n", " batch_size=torch.Size([32, 99]),\n", " device=None,\n", " is_shared=False),\n", " targets: TensorDict(\n", " fields={\n", " data: Tensor(shape=torch.Size([32, 99, 1]), device=cpu, dtype=torch.float32, is_shared=False)},\n", " batch_size=torch.Size([32, 99]),\n", " device=None,\n", " is_shared=False)},\n", " batch_size=torch.Size([32]),\n", " device=None,\n", " is_shared=False)" ] }, "execution_count": 9, "metadata": {}, "output_type": "execute_result" } ], "source": [ "next(iter(dataloaders[\"train\"]))" ] }, { "cell_type": "markdown", "id": "91d00b7d", "metadata": {}, "source": [ "## Model Definition" ] }, { "cell_type": "code", "execution_count": 10, "id": "2ccfb96c", "metadata": {}, "outputs": [], "source": [ "state_size = 32\n", "\n", "builder = wm.WorldMachineBuilder(state_size=state_size, \n", " max_context_size=99, \n", " positional_encoder_type=\"alibi\")\n", "\n", "builder.state_activation = \"tanh\"" ] }, { "cell_type": "markdown", "id": "31e82409", "metadata": {}, "source": [ "After defining the basic parameters, we need to define our sensory channels. In our case, the dataset has only one sensory channel, \"data\". We also define the encoders and decoders of this channel. The encoded channel will have a size of 10." ] }, { "cell_type": "code", "execution_count": 11, "id": "45d67293", "metadata": {}, "outputs": [], "source": [ "data_encoded_size = 10\n", "\n", "builder.add_sensory_channel(channel_name=\"data\", \n", " channel_size=data_encoded_size,\n", " encoder=wm.layers.PointwiseFeedforward(input_dim=1, hidden_size=2*state_size, output_dim=data_encoded_size),\n", " decoder=wm.layers.PointwiseFeedforward(input_dim=state_size, hidden_size=2*state_size, output_dim=1))" ] }, { "cell_type": "markdown", "id": "5fa5b36e", "metadata": {}, "source": [ "Then we can define our blocks. We are going to use two sensory blocks. One is conditioned on the \"data\" channel, and the other on the \"latent world state\" as it was at the start of the temporal step:" ] }, { "cell_type": "code", "execution_count": 12, "id": "5ca8c834", "metadata": {}, "outputs": [], "source": [ "builder.add_block(count=1,\n", " sensory_channel=\"data\",\n", " n_attention_head=2)\n", "\n", "builder.add_block(count=1,\n", " sensory_channel=\"state\",\n", " n_attention_head=2)\n" ] }, { "cell_type": "markdown", "id": "3bb4cd9f", "metadata": {}, "source": [ "Finally, we can build our model:\n" ] }, { "cell_type": "code", "execution_count": 13, "id": "3a996855", "metadata": {}, "outputs": [], "source": [ "model = builder.build()" ] }, { "cell_type": "markdown", "id": "f50b60cc", "metadata": {}, "source": [ "## Training" ] }, { "cell_type": "markdown", "id": "7dc3eb2c", "metadata": {}, "source": [ "The first step to training is defining our criterions. In this case, we are going to train using an MSE criterion computed on the predictions of our only \"data\" sensory channel:" ] }, { "cell_type": "code", "execution_count": 14, "id": "3a3dd648", "metadata": {}, "outputs": [], "source": [ "cs = wm.train.CriterionSet()\n", "\n", "cs.add_sensory_criterion(name=\"mse\",\n", " sensory_channel=\"data\",\n", " criterion=torch.nn.MSELoss(),\n", " train=True)" ] }, { "cell_type": "markdown", "id": "fa5452ae", "metadata": {}, "source": [ "Then, we define our training stages. They operate in callbacks during training, performing diverse operations. The first stages we are going to add are the `StateManager`, which implements the state discovery process and enables model training, and `SensoryMasker`, which allows the model to handle the absence of sensory data. Both are essential for the model:" ] }, { "cell_type": "code", "execution_count": 15, "id": "5d40a007", "metadata": {}, "outputs": [], "source": [ "stages = []\n", "\n", "stages.append(wm.train.stages.StateManager())\n", "stages.append(wm.train.stages.SensoryMasker(mask_percentage=wm.train.UniformScheduler(low_value=0, \n", " high_value=1, \n", " n_epoch=0)))\n" ] }, { "cell_type": "markdown", "id": "4ec6aec4", "metadata": {}, "source": [ "The following stages are `SequenceBreaker`, `ShortTimeRecaller`, and `NoiseAdder`. They perform operations that can improve the model's performance:" ] }, { "cell_type": "code", "execution_count": 16, "id": "5a7757c7", "metadata": {}, "outputs": [], "source": [ "stages.append(wm.train.stages.SequenceBreaker(n_segment=2, fast_forward=True))\n", "\n", "stages.append(wm.train.stages.ShortTimeRecaller(\n", " channel_sizes={\"data\":1},\n", " criterions={\"data\":torch.nn.MSELoss()},\n", " n_past=5,\n", " n_future=5,\n", " stride_past=3,\n", " stride_future=3\n", "))\n", "\n", "stages.append(wm.train.stages.NoiseAdder(\n", " means={\"data\":0.0}, \n", " stds={\"data\":0.1}, \n", " mins={\"data\":-1,},\n", " maxs={\"data\":1.0}\n", "))" ] }, { "cell_type": "markdown", "id": "8d7e5b4c", "metadata": {}, "source": [ "`EarlyStopper` saves an intermediary model at the best epoch. The final model will have the weights of this epoch. It doesn't interrupt the training." ] }, { "cell_type": "code", "execution_count": 17, "id": "676e80e2", "metadata": {}, "outputs": [], "source": [ "stages.append(wm.train.stages.EarlyStopper())" ] }, { "cell_type": "markdown", "id": "2aaf7ada", "metadata": {}, "source": [ "We will use AdamW as an optimizer. We will also use Cosine Annealing with Warmup to schedule our learning rate: " ] }, { "cell_type": "code", "execution_count": 18, "id": "7e7bac64", "metadata": {}, "outputs": [], "source": [ "initial_lr = 0.002\n", "T_mult = 1\n", "T0 = 5\n", "\n", "optimizer = torch.optim.AdamW(model.parameters(), lr=initial_lr)\n", "\n", "scheduler = torch.optim.lr_scheduler.CosineAnnealingWarmRestarts(optimizer, T0, T_mult)\n", "\n", "stages.append(wm.train.stages.LRScheduler_Step(scheduler=scheduler))" ] }, { "cell_type": "markdown", "id": "89286870", "metadata": {}, "source": [ "Finally, we can create our `Trainer` with the defined criterions and stages, and train our model for 20 epochs:" ] }, { "cell_type": "code", "execution_count": 19, "id": "b9238228", "metadata": {}, "outputs": [], "source": [ "trainer = wm.train.Trainer(criterion_set=cs, stages=stages)" ] }, { "cell_type": "code", "execution_count": 20, "id": "ddb44aa4", "metadata": {}, "outputs": [], "source": [ "model = model.to(device)" ] }, { "cell_type": "code", "execution_count": 21, "id": "2ee8d8da", "metadata": {}, "outputs": [ { "name": "stderr", "output_type": "stream", "text": [ "100%|██████████| 9/9 [00:00<00:00, 15.98it/s]\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "VAL Epoch [0/20], Loss: 0.3791\n" ] }, { "name": "stderr", "output_type": "stream", "text": [ "100%|██████████| 21/21 [00:03<00:00, 5.72it/s]\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "Epoch [1/20], Loss: 0.2650, Elapsed Time: 3.68 sec\n" ] }, { "name": "stderr", "output_type": "stream", "text": [ "100%|██████████| 9/9 [00:00<00:00, 44.05it/s]\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "VAL Epoch [1/20], Loss: 0.2643\n" ] }, { "name": "stderr", "output_type": "stream", "text": [ "100%|██████████| 21/21 [00:01<00:00, 17.74it/s]\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "Epoch [2/20], Loss: 0.2391, Elapsed Time: 1.19 sec\n" ] }, { "name": "stderr", "output_type": "stream", "text": [ "100%|██████████| 9/9 [00:00<00:00, 44.13it/s]\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "VAL Epoch [2/20], Loss: 0.1644\n" ] }, { "name": "stderr", "output_type": "stream", "text": [ "100%|██████████| 21/21 [00:01<00:00, 17.16it/s]\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "Epoch [3/20], Loss: 0.1699, Elapsed Time: 1.23 sec\n" ] }, { "name": "stderr", "output_type": "stream", "text": [ "100%|██████████| 9/9 [00:00<00:00, 44.54it/s]\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "VAL Epoch [3/20], Loss: 0.1028\n" ] }, { "name": "stderr", "output_type": "stream", "text": [ "100%|██████████| 21/21 [00:01<00:00, 16.54it/s]\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "Epoch [4/20], Loss: 0.1534, Elapsed Time: 1.28 sec\n" ] }, { "name": "stderr", "output_type": "stream", "text": [ "100%|██████████| 9/9 [00:00<00:00, 41.40it/s]\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "VAL Epoch [4/20], Loss: 0.1032\n" ] }, { "name": "stderr", "output_type": "stream", "text": [ "100%|██████████| 21/21 [00:01<00:00, 17.05it/s]\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "Epoch [5/20], Loss: 0.1435, Elapsed Time: 1.24 sec\n" ] }, { "name": "stderr", "output_type": "stream", "text": [ "100%|██████████| 9/9 [00:00<00:00, 38.23it/s]\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "VAL Epoch [5/20], Loss: 0.0870\n" ] }, { "name": "stderr", "output_type": "stream", "text": [ "100%|██████████| 21/21 [00:01<00:00, 17.49it/s]\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "Epoch [6/20], Loss: 0.1438, Elapsed Time: 1.21 sec\n" ] }, { "name": "stderr", "output_type": "stream", "text": [ "100%|██████████| 9/9 [00:00<00:00, 42.00it/s]\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "VAL Epoch [6/20], Loss: 0.0507\n" ] }, { "name": "stderr", "output_type": "stream", "text": [ "100%|██████████| 21/21 [00:01<00:00, 17.07it/s]\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "Epoch [7/20], Loss: 0.1223, Elapsed Time: 1.25 sec\n" ] }, { "name": "stderr", "output_type": "stream", "text": [ "100%|██████████| 9/9 [00:00<00:00, 37.87it/s]\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "VAL Epoch [7/20], Loss: 0.0499\n" ] }, { "name": "stderr", "output_type": "stream", "text": [ "100%|██████████| 21/21 [00:01<00:00, 17.52it/s]\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "Epoch [8/20], Loss: 0.1155, Elapsed Time: 1.21 sec\n" ] }, { "name": "stderr", "output_type": "stream", "text": [ "100%|██████████| 9/9 [00:00<00:00, 39.03it/s]\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "VAL Epoch [8/20], Loss: 0.0200\n" ] }, { "name": "stderr", "output_type": "stream", "text": [ "100%|██████████| 21/21 [00:01<00:00, 16.65it/s]\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "Epoch [9/20], Loss: 0.0986, Elapsed Time: 1.27 sec\n" ] }, { "name": "stderr", "output_type": "stream", "text": [ "100%|██████████| 9/9 [00:00<00:00, 44.25it/s]\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "VAL Epoch [9/20], Loss: 0.0154\n" ] }, { "name": "stderr", "output_type": "stream", "text": [ "100%|██████████| 21/21 [00:01<00:00, 16.83it/s]\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "Epoch [10/20], Loss: 0.0823, Elapsed Time: 1.26 sec\n" ] }, { "name": "stderr", "output_type": "stream", "text": [ "100%|██████████| 9/9 [00:00<00:00, 43.13it/s]\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "VAL Epoch [10/20], Loss: 0.0232\n" ] }, { "name": "stderr", "output_type": "stream", "text": [ "100%|██████████| 21/21 [00:01<00:00, 16.64it/s]\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "Epoch [11/20], Loss: 0.0756, Elapsed Time: 1.27 sec\n" ] }, { "name": "stderr", "output_type": "stream", "text": [ "100%|██████████| 9/9 [00:00<00:00, 45.19it/s]\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "VAL Epoch [11/20], Loss: 0.0255\n" ] }, { "name": "stderr", "output_type": "stream", "text": [ "100%|██████████| 21/21 [00:01<00:00, 17.38it/s]\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "Epoch [12/20], Loss: 0.0542, Elapsed Time: 1.22 sec\n" ] }, { "name": "stderr", "output_type": "stream", "text": [ "100%|██████████| 9/9 [00:00<00:00, 43.34it/s]\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "VAL Epoch [12/20], Loss: 0.0122\n" ] }, { "name": "stderr", "output_type": "stream", "text": [ "100%|██████████| 21/21 [00:01<00:00, 16.70it/s]\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "Epoch [13/20], Loss: 0.0495, Elapsed Time: 1.27 sec\n" ] }, { "name": "stderr", "output_type": "stream", "text": [ "100%|██████████| 9/9 [00:00<00:00, 43.40it/s]\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "VAL Epoch [13/20], Loss: 0.0097\n" ] }, { "name": "stderr", "output_type": "stream", "text": [ "100%|██████████| 21/21 [00:01<00:00, 18.14it/s]\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "Epoch [14/20], Loss: 0.0441, Elapsed Time: 1.17 sec\n" ] }, { "name": "stderr", "output_type": "stream", "text": [ "100%|██████████| 9/9 [00:00<00:00, 39.84it/s]\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "VAL Epoch [14/20], Loss: 0.0097\n" ] }, { "name": "stderr", "output_type": "stream", "text": [ "100%|██████████| 21/21 [00:01<00:00, 17.96it/s]\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "Epoch [15/20], Loss: 0.0403, Elapsed Time: 1.18 sec\n" ] }, { "name": "stderr", "output_type": "stream", "text": [ "100%|██████████| 9/9 [00:00<00:00, 47.19it/s]\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "VAL Epoch [15/20], Loss: 0.0090\n" ] }, { "name": "stderr", "output_type": "stream", "text": [ "100%|██████████| 21/21 [00:01<00:00, 18.79it/s]\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "Epoch [16/20], Loss: 0.0390, Elapsed Time: 1.13 sec\n" ] }, { "name": "stderr", "output_type": "stream", "text": [ "100%|██████████| 9/9 [00:00<00:00, 46.74it/s]\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "VAL Epoch [16/20], Loss: 0.0088\n" ] }, { "name": "stderr", "output_type": "stream", "text": [ "100%|██████████| 21/21 [00:01<00:00, 18.08it/s]\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "Epoch [17/20], Loss: 0.0304, Elapsed Time: 1.17 sec\n" ] }, { "name": "stderr", "output_type": "stream", "text": [ "100%|██████████| 9/9 [00:00<00:00, 47.40it/s]\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "VAL Epoch [17/20], Loss: 0.0104\n" ] }, { "name": "stderr", "output_type": "stream", "text": [ "100%|██████████| 21/21 [00:01<00:00, 18.33it/s]\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "Epoch [18/20], Loss: 0.0291, Elapsed Time: 1.15 sec\n" ] }, { "name": "stderr", "output_type": "stream", "text": [ "100%|██████████| 9/9 [00:00<00:00, 42.70it/s]\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "VAL Epoch [18/20], Loss: 0.0081\n" ] }, { "name": "stderr", "output_type": "stream", "text": [ "100%|██████████| 21/21 [00:01<00:00, 16.83it/s]\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "Epoch [19/20], Loss: 0.0266, Elapsed Time: 1.26 sec\n" ] }, { "name": "stderr", "output_type": "stream", "text": [ "100%|██████████| 9/9 [00:00<00:00, 44.56it/s]\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "VAL Epoch [19/20], Loss: 0.0072\n" ] }, { "name": "stderr", "output_type": "stream", "text": [ "100%|██████████| 21/21 [00:01<00:00, 17.83it/s]\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "Epoch [20/20], Loss: 0.0270, Elapsed Time: 1.19 sec\n" ] }, { "name": "stderr", "output_type": "stream", "text": [ "100%|██████████| 9/9 [00:00<00:00, 39.71it/s]" ] }, { "name": "stdout", "output_type": "stream", "text": [ "VAL Epoch [20/20], Loss: 0.0079\n" ] }, { "name": "stderr", "output_type": "stream", "text": [ "\n" ] } ], "source": [ "train_history = trainer(wm=model, \n", " dataloaders=dataloaders, \n", " optimizer=optimizer, \n", " n_epoch=20)" ] }, { "cell_type": "markdown", "id": "9a2ad812", "metadata": {}, "source": [ "We can see the training loss during training:" ] }, { "cell_type": "code", "execution_count": 22, "id": "2086b4e5", "metadata": {}, "outputs": [ { "data": { "text/plain": [ "Text(0.5, 1.0, 'Train History')" ] }, "execution_count": 22, "metadata": {}, "output_type": "execute_result" }, { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "plt.plot(train_history[\"optimizer_loss_train\"], \"o-\", label=\"Train\")\n", "plt.plot(train_history[\"optimizer_loss_val\"], \"o-\", label=\"Validation\")\n", "\n", "plt.xlabel(\"Epoch\")\n", "plt.ylabel(\"Optimizer Loss\")\n", "\n", "plt.legend()\n", "\n", "\n", "plt.title(\"Train History\")" ] }, { "cell_type": "markdown", "id": "d4d13778", "metadata": {}, "source": [ "Finally, let's ensure that only the intermediary model is deleted. In a typical script (not a Jupyter Notebook), this is not necessary:" ] }, { "cell_type": "code", "execution_count": null, "id": "00162e2b", "metadata": {}, "outputs": [], "source": [ "wm.train.stages.EarlyStopper._delete_files()" ] }, { "cell_type": "markdown", "id": "0ec9ada6", "metadata": {}, "source": [ "And also ensure that the states used during training are also deleted (again, only because we are running in a Jupyter Notebook):" ] }, { "cell_type": "code", "execution_count": null, "id": "b2003595", "metadata": {}, "outputs": [], "source": [ "train_dataset.clear_states()\n", "val_dataset.clear_states()" ] }, { "cell_type": "markdown", "id": "b0248d4c", "metadata": {}, "source": [ "## Inference" ] }, { "cell_type": "markdown", "id": "a3959374", "metadata": {}, "source": [ "For inferencing with the model, we need to define our initial state and use the `inference` method, passing in our sensory data as well. It is the \"inputs\" element of the dataloader item." ] }, { "cell_type": "code", "execution_count": 23, "id": "3f954273", "metadata": {}, "outputs": [], "source": [ "item = next(iter(dataloaders[\"val\"]))" ] }, { "cell_type": "code", "execution_count": 25, "id": "c972b82a", "metadata": {}, "outputs": [], "source": [ "state = torch.zeros([batch_size, 99, state_size], device=device)\n", "\n", "output = model.inference(state=state,\n", " sensory_data=item[\"inputs\"])" ] }, { "cell_type": "code", "execution_count": 26, "id": "85e0afc2", "metadata": {}, "outputs": [], "source": [ "data_pred = output[\"data\"].cpu().detach()" ] }, { "cell_type": "code", "execution_count": 28, "id": "306e22af", "metadata": {}, "outputs": [ { "data": { "text/plain": [ "Text(0.5, 1.0, 'Inference sample')" ] }, "execution_count": 28, "metadata": {}, "output_type": "execute_result" }, { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "plt.plot(item[\"targets\"][\"data\"][0])\n", "plt.plot(data_pred[0])\n", "\n", "plt.xlabel(\"Time\")\n", "plt.ylabel(\"Data\")\n", "plt.title(\"Inference sample\")" ] }, { "cell_type": "markdown", "id": "cfe53b2d", "metadata": {}, "source": [ "## Evaluating the model" ] }, { "cell_type": "markdown", "id": "b3bd73c9", "metadata": {}, "source": [ "We will evaluate our model's performance on the defined evaluation tasks. We start by creating our `MetricsGenerator` using our criterions:" ] }, { "cell_type": "code", "execution_count": 29, "id": "a214fea3", "metadata": {}, "outputs": [], "source": [ "mg = wm.evaluate.MetricsGenerator(cs)" ] }, { "cell_type": "markdown", "id": "f0658511", "metadata": {}, "source": [ "We can then compute the model metrics:" ] }, { "cell_type": "code", "execution_count": 30, "id": "f5df8d4b", "metadata": {}, "outputs": [ { "name": "stderr", "output_type": "stream", "text": [ "Metrics Generation: 100%|██████████| 9/9 [00:12<00:00, 1.36s/it]\n" ] } ], "source": [ "metrics = mg(model, dataloaders[\"val\"])" ] }, { "cell_type": "markdown", "id": "ad47c9e7", "metadata": {}, "source": [ "Plotting our metrics, we can see that the Prediction Shallow is one of the most challenging tasks, and Normal is the easiest:" ] }, { "cell_type": "code", "execution_count": 36, "id": "c55d1ef5", "metadata": {}, "outputs": [ { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "metrics_optimizer = {name.replace(\"_\", \"\\n\"):metrics[name][\"optimizer_loss\"] for name in metrics}\n", "\n", "sns.barplot(metrics_optimizer)\n", "plt.title(\"Performance Across Tasks\")\n", "plt.xlabel(\"Task\")\n", "plt.ylabel(\"MSE\")\n", "plt.yscale(\"log\")" ] }, { "cell_type": "markdown", "id": "48b49b6b", "metadata": {}, "source": [ "Let's also see some inference samples. For that, we are going to use a subsample of only one batch:" ] }, { "cell_type": "code", "execution_count": 73, "id": "79b53eb2", "metadata": {}, "outputs": [], "source": [ "small_dataset = val_dataset\n", "small_dataset, _ = torch.utils.data.random_split(small_dataset, [32, len(small_dataset)-32])\n", "\n", "small_dataloader = wm.data.WorldMachineDataLoader(small_dataset, batch_size=32, shuffle=False)" ] }, { "cell_type": "code", "execution_count": 74, "id": "9766ff60", "metadata": {}, "outputs": [ { "name": "stderr", "output_type": "stream", "text": [ "Metrics Generation: 0%| | 0/1 [00:00" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "plt.plot(logits[\"targets\"][\"data\"][i], label=\"Target\")\n", "plt.plot(logits[\"normal\"][\"data\"][i], label=\"Normal\")\n", "plt.plot(range(50,100), logits[\"prediction_shallow\"][\"data\"][i], label=\"Prediction Shallow\")\n", "\n", "plt.legend(bbox_to_anchor=(1.4, 1), loc='upper right')\n", "\n", "plt.title(\"Task Sample\")\n", "plt.xlabel(\"Time\")\n", "plt.ylabel(\"Data\")\n", "\n", "plt.show()" ] } ], "metadata": { "kernelspec": { "display_name": "Python 3", "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": "3.12.8" } }, "nbformat": 4, "nbformat_minor": 5 }