notebooks/MULT-MOSI.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "49e0ad50-a2a2-4b96-b360-6ebd4ba790f6",
   "metadata": {},
   "source": [
    "# Multimodal Transformer Model - MOSI"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1601d4a9-c176-4e94-9c84-250ae7ceb246",
   "metadata": {},
   "source": [
    "In this notebook, we implement the MulT Model (Multimodal Transformer for Unaligned Multimodal Language Sequences), and test it on CMU-MOSI dataset."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0e330111-2839-49fa-86f3-2204b6d24fc4",
   "metadata": {},
   "source": [
    "## Importing Libraries"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "90273986-226c-4f13-a783-27d03eda1e21",
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "import sys\n",
    "import torch\n",
    "\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "sys.path.append(os.getcwd())\n",
    "sys.path.append(os.path.dirname(os.path.dirname(os.getcwd())))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3a8511b4-49ef-4df0-916e-07c8002ebcf9",
   "metadata": {},
   "source": [
    "## Importing Helper Functions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "77dcaf0e-287e-4312-a80b-a67ca4221f73",
   "metadata": {},
   "outputs": [],
   "source": [
    "from helpers.data.get_data import get_dataloader\n",
    "from helpers.unimodals.common_models import Identity, MLP\n",
    "from helpers.fusions.mult import MULTModel\n",
    "from helpers.fusions.common_fusions import Concat\n",
    "from helpers.training.supervised_learning import train, test\n",
    "from helpers.utils.common_utils import get_default_device, save_results"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2bc745d1-4151-4894-af83-77780036ba7b",
   "metadata": {},
   "source": [
    "## Loading the Dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "c7aa2aea-6160-4f43-a6f1-596f9c26bab7",
   "metadata": {},
   "outputs": [],
   "source": [
    "traindata, validdata, testdata = get_dataloader('/scratch/rrm9598/data/mosi_data.pkl', robust_test=False, max_pad=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "3e83b4bf-2838-4c0b-beb8-3d6150a369f3",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "cuda\n"
     ]
    }
   ],
   "source": [
    "device = get_default_device()\n",
    "print(device)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "da5644c7-a89a-437f-8b75-3bf386b20dea",
   "metadata": {},
   "source": [
    "## Model Architecture"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "55070fb4-ea3e-4529-a717-7a0b3baf5275",
   "metadata": {},
   "outputs": [],
   "source": [
    "class HParams():\n",
    "    num_heads = 8\n",
    "    layers = 4\n",
    "    attn_dropout = 0.1\n",
    "    attn_dropout_modalities = [0, 0, 0.1]\n",
    "    relu_dropout = 0.1\n",
    "    res_dropout = 0.1\n",
    "    out_dropout = 0.1\n",
    "    embed_dropout = 0.2\n",
    "    embed_dim = 40\n",
    "    attn_mask = True\n",
    "    output_dim = 1\n",
    "    all_steps = False"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "b897d94f-f547-4100-aa74-80076e92e8fd",
   "metadata": {},
   "outputs": [],
   "source": [
    "encoders = [\n",
    "    Identity().to(device), \n",
    "    Identity().to(device), \n",
    "    Identity().to(device)\n",
    "]\n",
    "\n",
    "fusion = MULTModel(3, [20, 5, 300], hyp_params=HParams).to(device)\n",
    "\n",
    "head = Identity().to(device)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1ccb00f7-9da1-43bb-b41e-f39eca505a76",
   "metadata": {},
   "source": [
    "## Training"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "334d11e7-12b3-476b-8a3e-ebc39c814460",
   "metadata": {},
   "outputs": [],
   "source": [
    "FILE_NAME = 'mosi_mult_r0'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "49704ef0-fdb1-4f25-93a3-25083ce2d016",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch [1/100]\n",
      "Train Loss: 1.4333\n",
      "Valid Loss: 1.3701\n",
      "Saving Best\n",
      "\n",
      "Epoch [2/100]\n",
      "Train Loss: 1.1188\n",
      "Valid Loss: 1.2246\n",
      "Saving Best\n",
      "\n",
      "Epoch [3/100]\n",
      "Train Loss: 0.9708\n",
      "Valid Loss: 1.0582\n",
      "Saving Best\n",
      "\n",
      "Epoch [4/100]\n",
      "Train Loss: 0.9027\n",
      "Valid Loss: 1.0905\n",
      "\n",
      "Epoch [5/100]\n",
      "Train Loss: 0.8551\n",
      "Valid Loss: 1.0067\n",
      "Saving Best\n",
      "\n",
      "Epoch [6/100]\n",
      "Train Loss: 0.8109\n",
      "Valid Loss: 1.0238\n",
      "\n",
      "Epoch [7/100]\n",
      "Train Loss: 0.7728\n",
      "Valid Loss: 0.9285\n",
      "Saving Best\n",
      "\n",
      "Epoch [8/100]\n",
      "Train Loss: 0.7213\n",
      "Valid Loss: 1.0268\n",
      "\n",
      "Epoch [9/100]\n",
      "Train Loss: 0.7487\n",
      "Valid Loss: 1.1480\n",
      "\n",
      "Epoch [10/100]\n",
      "Train Loss: 0.7295\n",
      "Valid Loss: 0.9834\n",
      "\n",
      "Epoch [11/100]\n",
      "Train Loss: 0.6909\n",
      "Valid Loss: 0.9648\n",
      "\n",
      "Epoch [12/100]\n",
      "Train Loss: 0.6707\n",
      "Valid Loss: 1.0346\n",
      "\n",
      "Epoch [13/100]\n",
      "Train Loss: 0.6571\n",
      "Valid Loss: 0.9122\n",
      "Saving Best\n",
      "\n",
      "Epoch [14/100]\n",
      "Train Loss: 0.6843\n",
      "Valid Loss: 0.9369\n",
      "\n",
      "Epoch [15/100]\n",
      "Train Loss: 0.6172\n",
      "Valid Loss: 0.9335\n",
      "\n",
      "Epoch [16/100]\n",
      "Train Loss: 0.6266\n",
      "Valid Loss: 0.9503\n",
      "\n",
      "Epoch [17/100]\n",
      "Train Loss: 0.6109\n",
      "Valid Loss: 0.9586\n",
      "\n",
      "Epoch [18/100]\n",
      "Train Loss: 0.6349\n",
      "Valid Loss: 0.9544\n",
      "\n",
      "Epoch [19/100]\n",
      "Train Loss: 0.5737\n",
      "Valid Loss: 0.9742\n",
      "\n",
      "Epoch [20/100]\n",
      "Train Loss: 0.5735\n",
      "Valid Loss: 0.9344\n",
      "\n",
      "Epoch [21/100]\n",
      "Train Loss: 0.5513\n",
      "Valid Loss: 0.9508\n",
      "\n",
      "Epoch [22/100]\n",
      "Train Loss: 0.5594\n",
      "Valid Loss: 0.9589\n",
      "\n",
      "Epoch [23/100]\n",
      "Train Loss: 0.5421\n",
      "Valid Loss: 1.0112\n",
      "\n",
      "Epoch [24/100]\n",
      "Train Loss: 0.5272\n",
      "Valid Loss: 0.9654\n",
      "\n",
      "Epoch [25/100]\n",
      "Train Loss: 0.5232\n",
      "Valid Loss: 0.9693\n",
      "\n",
      "Epoch [26/100]\n",
      "Train Loss: 0.5259\n",
      "Valid Loss: 0.9971\n",
      "\n",
      "Epoch [27/100]\n",
      "Train Loss: 0.4953\n",
      "Valid Loss: 0.9873\n",
      "\n",
      "Epoch [28/100]\n",
      "Train Loss: 0.4839\n",
      "Valid Loss: 0.9343\n",
      "\n",
      "Epoch [29/100]\n",
      "Train Loss: 0.4710\n",
      "Valid Loss: 0.9365\n",
      "\n",
      "Epoch [30/100]\n",
      "Train Loss: 0.4593\n",
      "Valid Loss: 0.9452\n",
      "\n",
      "Epoch [31/100]\n",
      "Train Loss: 0.4541\n",
      "Valid Loss: 0.9592\n",
      "\n",
      "Epoch [32/100]\n",
      "Train Loss: 0.4595\n",
      "Valid Loss: 0.9848\n",
      "\n",
      "Epoch [33/100]\n",
      "Train Loss: 0.4730\n",
      "Valid Loss: 0.9608\n",
      "\n",
      "Epoch [34/100]\n",
      "Train Loss: 0.4246\n",
      "Valid Loss: 0.9638\n",
      "\n",
      "Epoch [35/100]\n",
      "Train Loss: 0.4444\n",
      "Valid Loss: 1.0026\n",
      "\n",
      "Epoch [36/100]\n",
      "Train Loss: 0.4181\n",
      "Valid Loss: 0.9447\n",
      "\n",
      "Epoch [37/100]\n",
      "Train Loss: 0.3901\n",
      "Valid Loss: 0.9517\n",
      "\n",
      "Epoch [38/100]\n",
      "Train Loss: 0.4000\n",
      "Valid Loss: 1.0195\n",
      "\n",
      "Epoch [39/100]\n",
      "Train Loss: 0.3979\n",
      "Valid Loss: 0.9928\n",
      "\n",
      "Epoch [40/100]\n",
      "Train Loss: 0.4016\n",
      "Valid Loss: 0.9430\n",
      "\n",
      "Epoch [41/100]\n",
      "Train Loss: 0.3877\n",
      "Valid Loss: 1.0325\n",
      "\n",
      "Epoch [42/100]\n",
      "Train Loss: 0.3802\n",
      "Valid Loss: 0.9911\n",
      "\n",
      "Epoch [43/100]\n",
      "Train Loss: 0.3919\n",
      "Valid Loss: 0.9833\n",
      "\n",
      "Epoch [44/100]\n",
      "Train Loss: 0.3582\n",
      "Valid Loss: 0.9835\n",
      "\n",
      "Epoch [45/100]\n",
      "Train Loss: 0.3710\n",
      "Valid Loss: 0.9889\n",
      "\n",
      "Epoch [46/100]\n",
      "Train Loss: 0.3732\n",
      "Valid Loss: 0.9742\n",
      "\n",
      "Epoch [47/100]\n",
      "Train Loss: 0.3597\n",
      "Valid Loss: 0.9482\n",
      "\n",
      "Epoch [48/100]\n",
      "Train Loss: 0.3628\n",
      "Valid Loss: 0.9956\n",
      "\n",
      "Epoch [49/100]\n",
      "Train Loss: 0.3637\n",
      "Valid Loss: 0.9502\n",
      "\n",
      "Epoch [50/100]\n",
      "Train Loss: 0.3406\n",
      "Valid Loss: 0.9680\n",
      "\n",
      "Epoch [51/100]\n",
      "Train Loss: 0.3695\n",
      "Valid Loss: 1.0249\n",
      "\n",
      "Epoch [52/100]\n",
      "Train Loss: 0.3786\n",
      "Valid Loss: 0.9389\n",
      "\n",
      "Epoch [53/100]\n",
      "Train Loss: 0.3273\n",
      "Valid Loss: 0.9945\n",
      "\n",
      "Epoch [54/100]\n",
      "Train Loss: 0.3355\n",
      "Valid Loss: 0.9385\n",
      "\n",
      "Epoch [55/100]\n",
      "Train Loss: 0.3336\n",
      "Valid Loss: 0.9649\n",
      "\n",
      "Epoch [56/100]\n",
      "Train Loss: 0.3272\n",
      "Valid Loss: 0.9998\n",
      "\n",
      "Epoch [57/100]\n",
      "Train Loss: 0.3179\n",
      "Valid Loss: 0.9636\n",
      "\n",
      "Epoch [58/100]\n",
      "Train Loss: 0.3063\n",
      "Valid Loss: 0.9699\n",
      "\n",
      "Epoch [59/100]\n",
      "Train Loss: 0.3104\n",
      "Valid Loss: 0.9547\n",
      "\n",
      "Epoch [60/100]\n",
      "Train Loss: 0.3418\n",
      "Valid Loss: 0.9501\n",
      "\n",
      "Epoch [61/100]\n",
      "Train Loss: 0.3089\n",
      "Valid Loss: 1.0300\n",
      "\n",
      "Epoch [62/100]\n",
      "Train Loss: 0.3088\n",
      "Valid Loss: 0.9378\n",
      "\n",
      "Epoch [63/100]\n",
      "Train Loss: 0.3164\n",
      "Valid Loss: 0.9450\n",
      "\n",
      "Epoch [64/100]\n",
      "Train Loss: 0.3006\n",
      "Valid Loss: 0.9937\n",
      "\n",
      "Epoch [65/100]\n",
      "Train Loss: 0.2867\n",
      "Valid Loss: 0.9571\n",
      "\n",
      "Epoch [66/100]\n",
      "Train Loss: 0.2872\n",
      "Valid Loss: 0.9996\n",
      "\n",
      "Epoch [67/100]\n",
      "Train Loss: 0.2882\n",
      "Valid Loss: 0.9506\n",
      "\n",
      "Epoch [68/100]\n",
      "Train Loss: 0.3117\n",
      "Valid Loss: 0.9597\n",
      "\n",
      "Epoch [69/100]\n",
      "Train Loss: 0.2936\n",
      "Valid Loss: 0.9575\n",
      "\n",
      "Epoch [70/100]\n",
      "Train Loss: 0.2743\n",
      "Valid Loss: 0.9282\n",
      "\n",
      "Epoch [71/100]\n",
      "Train Loss: 0.2848\n",
      "Valid Loss: 0.9252\n",
      "\n",
      "Epoch [72/100]\n",
      "Train Loss: 0.2702\n",
      "Valid Loss: 0.9342\n",
      "\n",
      "Epoch [73/100]\n",
      "Train Loss: 0.2659\n",
      "Valid Loss: 0.9488\n",
      "\n",
      "Epoch [74/100]\n",
      "Train Loss: 0.2840\n",
      "Valid Loss: 0.9410\n",
      "\n",
      "Epoch [75/100]\n",
      "Train Loss: 0.2715\n",
      "Valid Loss: 0.9693\n",
      "\n",
      "Epoch [76/100]\n",
      "Train Loss: 0.2690\n",
      "Valid Loss: 0.9635\n",
      "\n",
      "Epoch [77/100]\n",
      "Train Loss: 0.2706\n",
      "Valid Loss: 0.9461\n",
      "\n",
      "Epoch [78/100]\n",
      "Train Loss: 0.2620\n",
      "Valid Loss: 0.9585\n",
      "\n",
      "Epoch [79/100]\n",
      "Train Loss: 0.2640\n",
      "Valid Loss: 0.9777\n",
      "\n",
      "Epoch [80/100]\n",
      "Train Loss: 0.2821\n",
      "Valid Loss: 0.9479\n",
      "\n",
      "Epoch [81/100]\n",
      "Train Loss: 0.2613\n",
      "Valid Loss: 0.9911\n",
      "\n",
      "Epoch [82/100]\n",
      "Train Loss: 0.2514\n",
      "Valid Loss: 0.9377\n",
      "\n",
      "Epoch [83/100]\n",
      "Train Loss: 0.2495\n",
      "Valid Loss: 0.9451\n",
      "\n",
      "Epoch [84/100]\n",
      "Train Loss: 0.2597\n",
      "Valid Loss: 0.9992\n",
      "\n",
      "Epoch [85/100]\n",
      "Train Loss: 0.2648\n",
      "Valid Loss: 1.0057\n",
      "\n",
      "Epoch [86/100]\n",
      "Train Loss: 0.2635\n",
      "Valid Loss: 1.0148\n",
      "\n",
      "Epoch [87/100]\n",
      "Train Loss: 0.2410\n",
      "Valid Loss: 0.9487\n",
      "\n",
      "Epoch [88/100]\n",
      "Train Loss: 0.2480\n",
      "Valid Loss: 0.9788\n",
      "\n",
      "Epoch [89/100]\n",
      "Train Loss: 0.2342\n",
      "Valid Loss: 0.9337\n",
      "\n",
      "Epoch [90/100]\n",
      "Train Loss: 0.2617\n",
      "Valid Loss: 0.9343\n",
      "\n",
      "Epoch [91/100]\n",
      "Train Loss: 0.2320\n",
      "Valid Loss: 0.9708\n",
      "\n",
      "Epoch [92/100]\n",
      "Train Loss: 0.2434\n",
      "Valid Loss: 0.9732\n",
      "\n",
      "Epoch [93/100]\n",
      "Train Loss: 0.2383\n",
      "Valid Loss: 0.9480\n",
      "\n",
      "Epoch [94/100]\n",
      "Train Loss: 0.2489\n",
      "Valid Loss: 0.9285\n",
      "\n",
      "Epoch [95/100]\n",
      "Train Loss: 0.2269\n",
      "Valid Loss: 0.9637\n",
      "\n",
      "Epoch [96/100]\n",
      "Train Loss: 0.2324\n",
      "Valid Loss: 0.9337\n",
      "\n",
      "Epoch [97/100]\n",
      "Train Loss: 0.2332\n",
      "Valid Loss: 0.9439\n",
      "\n",
      "Epoch [98/100]\n",
      "Train Loss: 0.2315\n",
      "Valid Loss: 0.9386\n",
      "\n",
      "Epoch [99/100]\n",
      "Train Loss: 0.2324\n",
      "Valid Loss: 0.9975\n",
      "\n",
      "Epoch [100/100]\n",
      "Train Loss: 0.2250\n",
      "Valid Loss: 1.0192\n",
      "\n",
      "Training Time: 9m54s\n",
      "Training Peak Mem: 1350.1640625\n",
      "Training Params: 3076961\n"
     ]
    }
   ],
   "source": [
    "train_losses, valid_losses = train(\n",
    "    encoders, fusion, head, traindata, validdata, 100, \n",
    "    task=\"regression\", optimtype=torch.optim.AdamW, early_stop=False, is_packed=False, \n",
    "    lr=1e-3, clip_val=1.0, save=f'/scratch/rrm9598/Multimodal-Sentiment-Analysis/models/{FILE_NAME}.pt', \n",
    "    weight_decay=0.01, objective=torch.nn.L1Loss())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "6322ee5b-0442-47f3-a910-2533fee34036",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 1.0, 'Loss vs. No. of epochs')"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(train_losses, label='Train Loss')\n",
    "plt.plot(valid_losses, label='Validation Loss')\n",
    "plt.xlabel('Epochs')\n",
    "plt.ylabel('Loss')\n",
    "plt.legend()\n",
    "plt.grid(True)\n",
    "plt.title('Loss vs. No. of epochs')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "4ecbf0dc-8fad-490b-bd2a-5b0c265e91ab",
   "metadata": {},
   "outputs": [],
   "source": [
    "save_results(train_losses, valid_losses, f'/scratch/rrm9598/Multimodal-Sentiment-Analysis/results/{FILE_NAME}.pkl')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7c091121-ff36-40fc-ab64-a58a4f524e8b",
   "metadata": {},
   "source": [
    "## Testing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "d01b591b-614f-433c-9ae7-10ac18a850fe",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Recall: 76.6768% | Total Accuracy: 75.0729%\n",
      "Inference Time: 1s484ms\n",
      "Inference Params: 3076961\n"
     ]
    }
   ],
   "source": [
    "model = torch.load(f'/scratch/rrm9598/Multimodal-Sentiment-Analysis/models/{FILE_NAME}.pt').to(device)\n",
    "\n",
    "test(model, test_dataloaders_all=testdata, dataset='mosi', is_packed=False,\n",
    "     criterion=torch.nn.L1Loss(), task='posneg-classification', no_robust=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "36248b8b-cf37-43b3-bdda-f2d7bf3ecce0",
   "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": "3.11.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}