notebooks/BayesianModels_And_Custom_Pyro_Modules.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "import sys\n",
    "sys.path.insert(0,'..')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/ypq/anaconda3/envs/GD/lib/python3.9/site-packages/torchvision/io/image.py:13: UserWarning: Failed to load image Python extension: /home/ypq/anaconda3/envs/GD/lib/python3.9/site-packages/torchvision/image.so: undefined symbol: _ZN5torch3jit17parseSchemaOrNameERKNSt7__cxx1112basic_stringIcSt11char_traitsIcESaIcEEE\n",
      "  warn(f\"Failed to load image Python extension: {e}\")\n"
     ]
    }
   ],
   "source": [
    "import torch\n",
    "import numpy as np\n",
    "\n",
    "import priors\n",
    "from train import train, get_weighted_single_eval_pos_sampler\n",
    "import encoders\n",
    "import positional_encodings\n",
    "import utils\n",
    "import bar_distribution\n",
    "import decoders\n",
    "from datasets import *\n",
    "import os\n",
    "\n",
    "from tqdm import tqdm\n",
    "import time\n",
    "\n",
    "import torch\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import pyro.distributions as dist\n",
    "\n",
    "import torch.nn as nn\n",
    "import os.path\n",
    "import glob\n",
    "\n",
    "from mcmc_svi_transformer_on_bayesian import get_model, get_default_model_spec, generate_toy_data, load_results, plot_with_confidence_intervals, training_steps, training_samples, get_default_evaluation_points, compute_mean_and_conf_interval, eval_transformer\n",
    "from pyro.nn import PyroModule, PyroParam, PyroSample\n",
    "import pyro\n",
    "import pyro.contrib.gp as gp\n",
    "import pyro.distributions as dist\n",
    "import pdb\n",
    "import scienceplots"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "%load_ext autoreload\n",
    "\n",
    "%autoreload 2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "## DEFINE A PRIOR MODEL ##\n",
    "# We define a Bayesian Model as a prior for all methods\n",
    "# This can be replaced by other models that inherit from PyroModule.\n",
    "class BayesianModel(PyroModule):\n",
    "    def __init__(self, model_spec, device='cuda'):\n",
    "        super().__init__()\n",
    "\n",
    "        self.device = device\n",
    "        self.num_features = model_spec['num_features']\n",
    "\n",
    "        mu, sigma = torch.tensor([0.0]).to(self.device), torch.tensor([1.0]).to(self.device)\n",
    "\n",
    "        self.fc1 = PyroModule[nn.Linear](self.num_features, model_spec['embed'])\n",
    "        self.fc1.weight = PyroSample(\n",
    "            dist.Normal(mu, sigma).expand([model_spec['embed'], self.num_features]).to_event(2))\n",
    "        self.fc1.bias = PyroSample(dist.Normal(mu, sigma).expand([model_spec['embed']]).to_event(1))\n",
    "\n",
    "        self.fc2 = PyroModule[nn.Linear](model_spec['embed'], 2)\n",
    "        self.fc2.weight = PyroSample(dist.Normal(mu, sigma).expand([2, model_spec['embed']]).to_event(2))\n",
    "        self.fc2.bias = PyroSample(dist.Normal(mu, sigma).expand([2]).to_event(1))\n",
    "\n",
    "        self.model = torch.nn.Sequential(self.fc1, self.fc2)\n",
    "\n",
    "        self.to(self.device)\n",
    "\n",
    "    def forward(self, x=None, y=None, seq_len=1):\n",
    "        if x is None:\n",
    "            with pyro.plate(\"x_plate\", seq_len):\n",
    "                d_ = dist.Normal(torch.tensor([0.0]).to(self.device), torch.tensor([1.0]).to(self.device)).expand(\n",
    "                    [self.num_features]).to_event(1) # sample uniformly from unit cube\n",
    "                x = pyro.sample(\"x\", d_)\n",
    "        out = self.model(x) # [seq_len,feature]\n",
    "        mu = out.squeeze() # [seq_len,feature]\n",
    "        softmax = torch.nn.Softmax(dim=1)\n",
    "        with pyro.plate(\"data\", out.shape[0]):\n",
    "            s = softmax(mu) # [seq_len,feature]\n",
    "            obs = pyro.sample('obs', dist.Categorical(probs=s), obs=y).float() # doing categoration, here 0 or 1\n",
    "\n",
    "        return x, obs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch\n",
    "import pyro\n",
    "import pyro.distributions as dist\n",
    "from pyro.nn import PyroModule, PyroSample\n",
    "\n",
    "class GPModel(PyroModule):\n",
    "    def __init__(self, model_spec, device='cuda'):\n",
    "        super().__init__()\n",
    "        self.device =device\n",
    "        self.num_features = model_spec['num_features']\n",
    "        self.kernel = gp.kernels.RBF(input_dim=self.num_features, variance=torch.tensor(1.),\n",
    "                        lengthscale=torch.tensor(.6))\n",
    "#         self.gpr = gp.models.GPRegression(X, y, kernel, noise=torch.tensor(1.))\n",
    "        self.to(self.device)\n",
    "\n",
    "    def forward(self, x=None, y=None, seq_len=1):\n",
    "        if x is None:\n",
    "            with pyro.plate(\"x_plate\", seq_len):\n",
    "                d_ = dist.Normal(torch.tensor([0.0]).to(self.device), torch.tensor([1.0]).to(self.device)).expand(\n",
    "                    [self.num_features]).to_event(1) # sample uniformly from unit cube\n",
    "                x = pyro.sample(\"x\", d_)\n",
    "        n = x.shape[0]\n",
    "        with pyro.plate(\"data\", 1):\n",
    "            obs = pyro.sample(\"obs\", dist.MultivariateNormal(torch.zeros(n).to(self.device), self.kernel(x).to(self.device)), obs=y).float()\n",
    "        return x,obs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "results_directory = '../results' # Where to save results\n",
    "model_spec_size = 'big' # Size of the BNN model to evaluate, also try big\n",
    "bptt = 100 # Number of samples in each dataset\n",
    "\n",
    "# Training samples seen after which to evaluate the methods\n",
    "evaluation_points = [2, 7, 12, 17, 22, 27, 32, 37, 42, 47, 52, 57, 62, 67, 72, 77, 82, 87, 92]\n",
    "model_spec = get_default_model_spec(model_spec_size)\n",
    "\n",
    "# Function which generates a model from the prior\n",
    "model_sampler = lambda : BayesianModel(model_spec, device = device)\n",
    "# model_sampler = lambda : GPModel(model_spec, device = device)\n",
    "\n",
    "global_results = {} # Dict in which to save results\n",
    "task = 'samples' # Task to evaluate, only option is samples, keep fixed"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "mkdir: cannot create directory ‘../results’: File exists\r\n"
     ]
    }
   ],
   "source": [
    "!mkdir {results_directory}"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "heading_collapsed": true
   },
   "source": [
    "### Evaluate SVI and MCMC"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "hidden": true
   },
   "outputs": [],
   "source": [
    "method = 'mcmc'\n",
    "steps = 1\n",
    "device = 'cpu'\n",
    "path_interfix = f'{results_directory}/timing_{model_spec_size}_model_test'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "test_model = model_sampler()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "hidden": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "mkdir: cannot create directory ‘../results/timing_big_model_test’: File exists\r\n"
     ]
    }
   ],
   "source": [
    "!mkdir {path_interfix}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "hidden": true,
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "already done 2\n",
      "already done 7\n",
      "already done 12\n",
      "already done 17\n",
      "already done 22\n",
      "already done 27\n",
      "already done 32\n",
      "already done 37\n",
      "already done 42\n",
      "already done 47\n",
      "already done 52\n",
      "already done 57\n",
      "already done 62\n",
      "already done 67\n",
      "already done 72\n",
      "already done 77\n",
      "already done 82\n",
      "already done 87\n",
      "already done 92\n"
     ]
    }
   ],
   "source": [
    "# This evaluates baseline method (SVI or MCMC) with varying number of trainings samples and steps number of inference steps\n",
    "\n",
    "X, y = generate_toy_data(test_model, bptt, device) \n",
    "# X.shape:[100,100,3](num_of_datasets,num_of_points_in_each_dataset,feature) Y.shape:[100,100]\n",
    "\n",
    "training_samples(method\n",
    "                 , X\n",
    "                 , y\n",
    "                 , model_sampler\n",
    "                 , evaluation_points\n",
    "                 , steps=steps\n",
    "                 , device=device\n",
    "                 , path_interfix=path_interfix)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "hidden": true
   },
   "outputs": [],
   "source": [
    "device = 'cuda'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "hidden": true
   },
   "outputs": [],
   "source": [
    "config = {'lr': 2.006434218345026e-05\n",
    " , 'epochs': 160\n",
    " , 'dropout': 0.0\n",
    " , 'emsize': 256\n",
    " , 'batch_size': 256\n",
    " , 'nlayers': 5\n",
    " , 'num_outputs': 1\n",
    " , 'num_features': model_spec['num_features']\n",
    " , 'steps_per_epoch': 100\n",
    " , 'nhead': 4\n",
    " , 'seq_len': model_spec['seq_len']\n",
    " , 'nhid_factor': 2}"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Training Transformer on Prior (Skip this step to reuse results)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "metadata": {
    "hidden": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Using cpu:0 device\n",
      "DataLoader.__dict__ {'num_steps': 100, 'fuse_x_y': False, 'get_batch_kwargs': {'batch_size': 256, 'seq_len': 300, 'num_outputs': 1, 'num_features': 3, 'canonical_args': None, 'model': <function <lambda> at 0x7fc4801af9d0>}, 'num_features': 3, 'num_outputs': 1}\n"
     ]
    },
    {
     "ename": "KeyboardInterrupt",
     "evalue": "",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m                         Traceback (most recent call last)",
      "Cell \u001b[0;32mIn[84], line 1\u001b[0m\n\u001b[0;32m----> 1\u001b[0m transformer_model \u001b[38;5;241m=\u001b[39m \u001b[43mget_model\u001b[49m\u001b[43m(\u001b[49m\u001b[43mmodel_sampler\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mconfig\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mshould_train\u001b[49m\u001b[43m \u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43m \u001b[49m\u001b[38;5;28;43;01mTrue\u001b[39;49;00m\u001b[43m)\u001b[49m\n\u001b[1;32m      2\u001b[0m model_path \u001b[38;5;241m=\u001b[39m os\u001b[38;5;241m.\u001b[39mpath\u001b[38;5;241m.\u001b[39mjoin(results_directory, \u001b[38;5;124mf\u001b[39m\u001b[38;5;124m'\u001b[39m\u001b[38;5;124mbayesian_models_transformer_checkpoint_\u001b[39m\u001b[38;5;132;01m{\u001b[39;00mmodel_spec_size\u001b[38;5;132;01m}\u001b[39;00m\u001b[38;5;124m_epochs_\u001b[39m\u001b[38;5;124m'\u001b[39m\u001b[38;5;241m+\u001b[39mconfig[\u001b[38;5;124m'\u001b[39m\u001b[38;5;124mepochs\u001b[39m\u001b[38;5;124m'\u001b[39m]\u001b[38;5;241m+\u001b[39m\u001b[38;5;124m'\u001b[39m\u001b[38;5;124m.cpkt\u001b[39m\u001b[38;5;124m'\u001b[39m)\n\u001b[1;32m      3\u001b[0m torch\u001b[38;5;241m.\u001b[39msave((transformer_model[\u001b[38;5;241m2\u001b[39m]\u001b[38;5;241m.\u001b[39mstate_dict(), \u001b[38;5;28;01mNone\u001b[39;00m), model_path)\n",
      "File \u001b[0;32m~/TransformersCanDoBayesianInference/notebooks/../mcmc_svi_transformer_on_bayesian.py:89\u001b[0m, in \u001b[0;36mget_model\u001b[0;34m(model_generator, config, should_train, device)\u001b[0m\n\u001b[1;32m     86\u001b[0m \u001b[38;5;28;01mdef\u001b[39;00m \u001b[38;5;21mget_model\u001b[39m(model_generator, config, should_train\u001b[38;5;241m=\u001b[39m\u001b[38;5;28;01mTrue\u001b[39;00m, device\u001b[38;5;241m=\u001b[39m\u001b[38;5;124m'\u001b[39m\u001b[38;5;124mcuda\u001b[39m\u001b[38;5;124m'\u001b[39m):\n\u001b[1;32m     87\u001b[0m     epochs \u001b[38;5;241m=\u001b[39m \u001b[38;5;241m0\u001b[39m \u001b[38;5;28;01mif\u001b[39;00m \u001b[38;5;129;01mnot\u001b[39;00m should_train \u001b[38;5;28;01melse\u001b[39;00m config[\u001b[38;5;124m'\u001b[39m\u001b[38;5;124mepochs\u001b[39m\u001b[38;5;124m'\u001b[39m]\n\u001b[0;32m---> 89\u001b[0m     model \u001b[38;5;241m=\u001b[39m \u001b[43mtrain\u001b[49m\u001b[43m(\u001b[49m\u001b[43mpriors\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mpyro\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mDataLoader\u001b[49m\n\u001b[1;32m     90\u001b[0m \u001b[43m                  \u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mLosses\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mbce\u001b[49m\n\u001b[1;32m     91\u001b[0m \u001b[43m                  \u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mencoders\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mLinear\u001b[49m\n\u001b[1;32m     92\u001b[0m \u001b[43m                  \u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43memsize\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43mconfig\u001b[49m\u001b[43m[\u001b[49m\u001b[38;5;124;43m'\u001b[39;49m\u001b[38;5;124;43memsize\u001b[39;49m\u001b[38;5;124;43m'\u001b[39;49m\u001b[43m]\u001b[49m\n\u001b[1;32m     93\u001b[0m \u001b[43m                  \u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mnhead\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43mconfig\u001b[49m\u001b[43m[\u001b[49m\u001b[38;5;124;43m'\u001b[39;49m\u001b[38;5;124;43mnhead\u001b[39;49m\u001b[38;5;124;43m'\u001b[39;49m\u001b[43m]\u001b[49m\n\u001b[1;32m     94\u001b[0m \u001b[43m                  \u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43my_encoder_generator\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43mencoders\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mLinear\u001b[49m\n\u001b[1;32m     95\u001b[0m \u001b[43m                  \u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mpos_encoder_generator\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[38;5;28;43;01mNone\u001b[39;49;00m\n\u001b[1;32m     96\u001b[0m \u001b[43m                  \u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mbatch_size\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43mconfig\u001b[49m\u001b[43m[\u001b[49m\u001b[38;5;124;43m'\u001b[39;49m\u001b[38;5;124;43mbatch_size\u001b[39;49m\u001b[38;5;124;43m'\u001b[39;49m\u001b[43m]\u001b[49m\n\u001b[1;32m     97\u001b[0m \u001b[43m                  \u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mnlayers\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43mconfig\u001b[49m\u001b[43m[\u001b[49m\u001b[38;5;124;43m'\u001b[39;49m\u001b[38;5;124;43mnlayers\u001b[39;49m\u001b[38;5;124;43m'\u001b[39;49m\u001b[43m]\u001b[49m\n\u001b[1;32m     98\u001b[0m \u001b[43m                  \u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mnhid\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43mconfig\u001b[49m\u001b[43m[\u001b[49m\u001b[38;5;124;43m'\u001b[39;49m\u001b[38;5;124;43memsize\u001b[39;49m\u001b[38;5;124;43m'\u001b[39;49m\u001b[43m]\u001b[49m\u001b[43m \u001b[49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[43m \u001b[49m\u001b[43mconfig\u001b[49m\u001b[43m[\u001b[49m\u001b[38;5;124;43m'\u001b[39;49m\u001b[38;5;124;43mnhid_factor\u001b[39;49m\u001b[38;5;124;43m'\u001b[39;49m\u001b[43m]\u001b[49m\n\u001b[1;32m     99\u001b[0m \u001b[43m                  \u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mepochs\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43mepochs\u001b[49m\n\u001b[1;32m    100\u001b[0m \u001b[43m                  \u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mwarmup_epochs\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43mconfig\u001b[49m\u001b[43m[\u001b[49m\u001b[38;5;124;43m'\u001b[39;49m\u001b[38;5;124;43mepochs\u001b[39;49m\u001b[38;5;124;43m'\u001b[39;49m\u001b[43m]\u001b[49m\u001b[43m \u001b[49m\u001b[38;5;241;43m/\u001b[39;49m\u001b[38;5;241;43m/\u001b[39;49m\u001b[43m \u001b[49m\u001b[38;5;241;43m4\u001b[39;49m\n\u001b[1;32m    101\u001b[0m \u001b[43m                  \u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mbptt\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43mconfig\u001b[49m\u001b[43m[\u001b[49m\u001b[38;5;124;43m'\u001b[39;49m\u001b[38;5;124;43mseq_len\u001b[39;49m\u001b[38;5;124;43m'\u001b[39;49m\u001b[43m]\u001b[49m\n\u001b[1;32m    102\u001b[0m \u001b[43m                  \u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mgpu_device\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43mdevice\u001b[49m\n\u001b[1;32m    103\u001b[0m \u001b[43m                  \u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mdropout\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43mconfig\u001b[49m\u001b[43m[\u001b[49m\u001b[38;5;124;43m'\u001b[39;49m\u001b[38;5;124;43mdropout\u001b[39;49m\u001b[38;5;124;43m'\u001b[39;49m\u001b[43m]\u001b[49m\n\u001b[1;32m    104\u001b[0m \u001b[43m                  \u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43msteps_per_epoch\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43mconfig\u001b[49m\u001b[43m[\u001b[49m\u001b[38;5;124;43m'\u001b[39;49m\u001b[38;5;124;43msteps_per_epoch\u001b[39;49m\u001b[38;5;124;43m'\u001b[39;49m\u001b[43m]\u001b[49m\n\u001b[1;32m    105\u001b[0m \u001b[43m                  \u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43msingle_eval_pos_gen\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43mget_weighted_single_eval_pos_sampler\u001b[49m\u001b[43m(\u001b[49m\u001b[38;5;241;43m100\u001b[39;49m\u001b[43m)\u001b[49m\n\u001b[1;32m    106\u001b[0m \u001b[43m                  \u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mextra_prior_kwargs_dict\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43m{\u001b[49m\n\u001b[1;32m    107\u001b[0m \u001b[43m            \u001b[49m\u001b[38;5;124;43m'\u001b[39;49m\u001b[38;5;124;43mnum_outputs\u001b[39;49m\u001b[38;5;124;43m'\u001b[39;49m\u001b[43m:\u001b[49m\u001b[43m \u001b[49m\u001b[43mconfig\u001b[49m\u001b[43m[\u001b[49m\u001b[38;5;124;43m'\u001b[39;49m\u001b[38;5;124;43mnum_outputs\u001b[39;49m\u001b[38;5;124;43m'\u001b[39;49m\u001b[43m]\u001b[49m\n\u001b[1;32m    108\u001b[0m \u001b[43m            \u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[38;5;124;43m'\u001b[39;49m\u001b[38;5;124;43mnum_features\u001b[39;49m\u001b[38;5;124;43m'\u001b[39;49m\u001b[43m:\u001b[49m\u001b[43m \u001b[49m\u001b[43mconfig\u001b[49m\u001b[43m[\u001b[49m\u001b[38;5;124;43m'\u001b[39;49m\u001b[38;5;124;43mnum_features\u001b[39;49m\u001b[38;5;124;43m'\u001b[39;49m\u001b[43m]\u001b[49m\n\u001b[1;32m    109\u001b[0m \u001b[43m            \u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[38;5;124;43m'\u001b[39;49m\u001b[38;5;124;43mcanonical_args\u001b[39;49m\u001b[38;5;124;43m'\u001b[39;49m\u001b[43m:\u001b[49m\u001b[43m \u001b[49m\u001b[38;5;28;43;01mNone\u001b[39;49;00m\n\u001b[1;32m    110\u001b[0m \u001b[43m            \u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[38;5;124;43m'\u001b[39;49m\u001b[38;5;124;43mfuse_x_y\u001b[39;49m\u001b[38;5;124;43m'\u001b[39;49m\u001b[43m:\u001b[49m\u001b[43m \u001b[49m\u001b[38;5;28;43;01mFalse\u001b[39;49;00m\n\u001b[1;32m    111\u001b[0m \u001b[43m            \u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[38;5;124;43m'\u001b[39;49m\u001b[38;5;124;43mmodel\u001b[39;49m\u001b[38;5;124;43m'\u001b[39;49m\u001b[43m:\u001b[49m\u001b[43m \u001b[49m\u001b[43mmodel_generator\u001b[49m\n\u001b[1;32m    112\u001b[0m \u001b[43m        \u001b[49m\u001b[43m}\u001b[49m\n\u001b[1;32m    113\u001b[0m \u001b[43m                  \u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mlr\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43mconfig\u001b[49m\u001b[43m[\u001b[49m\u001b[38;5;124;43m'\u001b[39;49m\u001b[38;5;124;43mlr\u001b[39;49m\u001b[38;5;124;43m'\u001b[39;49m\u001b[43m]\u001b[49m\n\u001b[1;32m    114\u001b[0m \u001b[43m                  \u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mverbose\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[38;5;28;43;01mTrue\u001b[39;49;00m\u001b[43m)\u001b[49m\n\u001b[1;32m    116\u001b[0m     \u001b[38;5;28;01mreturn\u001b[39;00m model\n",
      "File \u001b[0;32m~/TransformersCanDoBayesianInference/notebooks/../train.py:118\u001b[0m, in \u001b[0;36mtrain\u001b[0;34m(priordataloader_class, criterion, encoder_generator, emsize, nhid, nlayers, nhead, dropout, epochs, steps_per_epoch, batch_size, bptt, lr, warmup_epochs, input_normalization, y_encoder_generator, pos_encoder_generator, decoder, extra_prior_kwargs_dict, scheduler, load_weights_from_this_state_dict, validation_period, single_eval_pos_gen, gpu_device, aggregate_k_gradients, verbose)\u001b[0m\n\u001b[1;32m    116\u001b[0m \u001b[38;5;28;01mfor\u001b[39;00m epoch \u001b[38;5;129;01min\u001b[39;00m \u001b[38;5;28mrange\u001b[39m(\u001b[38;5;241m1\u001b[39m, epochs \u001b[38;5;241m+\u001b[39m \u001b[38;5;241m1\u001b[39m):\n\u001b[1;32m    117\u001b[0m     epoch_start_time \u001b[38;5;241m=\u001b[39m time\u001b[38;5;241m.\u001b[39mtime()\n\u001b[0;32m--> 118\u001b[0m     total_loss, total_positional_losses, time_to_get_batch, forward_time, step_time \u001b[38;5;241m=\u001b[39m \u001b[43mtrain\u001b[49m\u001b[43m(\u001b[49m\u001b[43m)\u001b[49m\n\u001b[1;32m    119\u001b[0m     \u001b[38;5;28;01mif\u001b[39;00m \u001b[38;5;28mhasattr\u001b[39m(dl, \u001b[38;5;124m'\u001b[39m\u001b[38;5;124mvalidate\u001b[39m\u001b[38;5;124m'\u001b[39m) \u001b[38;5;129;01mand\u001b[39;00m epoch \u001b[38;5;241m%\u001b[39m validation_period \u001b[38;5;241m==\u001b[39m \u001b[38;5;241m0\u001b[39m:\n\u001b[1;32m    120\u001b[0m         \u001b[38;5;28;01mwith\u001b[39;00m torch\u001b[38;5;241m.\u001b[39mno_grad():\n",
      "File \u001b[0;32m~/TransformersCanDoBayesianInference/notebooks/../train.py:93\u001b[0m, in \u001b[0;36mtrain.<locals>.train\u001b[0;34m()\u001b[0m\n\u001b[1;32m     89\u001b[0m losses \u001b[38;5;241m=\u001b[39m losses\u001b[38;5;241m.\u001b[39mview(\u001b[38;5;241m*\u001b[39moutput\u001b[38;5;241m.\u001b[39mshape[\u001b[38;5;241m0\u001b[39m:\u001b[38;5;241m2\u001b[39m])\u001b[38;5;241m.\u001b[39msqueeze(\u001b[38;5;241m-\u001b[39m\u001b[38;5;241m1\u001b[39m)\n\u001b[1;32m     92\u001b[0m loss \u001b[38;5;241m=\u001b[39m losses\u001b[38;5;241m.\u001b[39mmean()\n\u001b[0;32m---> 93\u001b[0m \u001b[43mloss\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mbackward\u001b[49m\u001b[43m(\u001b[49m\u001b[43m)\u001b[49m\n\u001b[1;32m     94\u001b[0m \u001b[38;5;28;01mif\u001b[39;00m batch \u001b[38;5;241m%\u001b[39m aggregate_k_gradients \u001b[38;5;241m==\u001b[39m aggregate_k_gradients \u001b[38;5;241m-\u001b[39m \u001b[38;5;241m1\u001b[39m:\n\u001b[1;32m     95\u001b[0m     torch\u001b[38;5;241m.\u001b[39mnn\u001b[38;5;241m.\u001b[39mutils\u001b[38;5;241m.\u001b[39mclip_grad_norm_(model\u001b[38;5;241m.\u001b[39mparameters(), \u001b[38;5;241m1.\u001b[39m)\n",
      "File \u001b[0;32m~/anaconda3/envs/GD/lib/python3.9/site-packages/torch/_tensor.py:488\u001b[0m, in \u001b[0;36mTensor.backward\u001b[0;34m(self, gradient, retain_graph, create_graph, inputs)\u001b[0m\n\u001b[1;32m    478\u001b[0m \u001b[38;5;28;01mif\u001b[39;00m has_torch_function_unary(\u001b[38;5;28mself\u001b[39m):\n\u001b[1;32m    479\u001b[0m     \u001b[38;5;28;01mreturn\u001b[39;00m handle_torch_function(\n\u001b[1;32m    480\u001b[0m         Tensor\u001b[38;5;241m.\u001b[39mbackward,\n\u001b[1;32m    481\u001b[0m         (\u001b[38;5;28mself\u001b[39m,),\n\u001b[0;32m   (...)\u001b[0m\n\u001b[1;32m    486\u001b[0m         inputs\u001b[38;5;241m=\u001b[39minputs,\n\u001b[1;32m    487\u001b[0m     )\n\u001b[0;32m--> 488\u001b[0m \u001b[43mtorch\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mautograd\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mbackward\u001b[49m\u001b[43m(\u001b[49m\n\u001b[1;32m    489\u001b[0m \u001b[43m    \u001b[49m\u001b[38;5;28;43mself\u001b[39;49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mgradient\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mretain_graph\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mcreate_graph\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43minputs\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43minputs\u001b[49m\n\u001b[1;32m    490\u001b[0m \u001b[43m\u001b[49m\u001b[43m)\u001b[49m\n",
      "File \u001b[0;32m~/anaconda3/envs/GD/lib/python3.9/site-packages/torch/autograd/__init__.py:197\u001b[0m, in \u001b[0;36mbackward\u001b[0;34m(tensors, grad_tensors, retain_graph, create_graph, grad_variables, inputs)\u001b[0m\n\u001b[1;32m    192\u001b[0m     retain_graph \u001b[38;5;241m=\u001b[39m create_graph\n\u001b[1;32m    194\u001b[0m \u001b[38;5;66;03m# The reason we repeat same the comment below is that\u001b[39;00m\n\u001b[1;32m    195\u001b[0m \u001b[38;5;66;03m# some Python versions print out the first line of a multi-line function\u001b[39;00m\n\u001b[1;32m    196\u001b[0m \u001b[38;5;66;03m# calls in the traceback and some print out the last line\u001b[39;00m\n\u001b[0;32m--> 197\u001b[0m \u001b[43mVariable\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43m_execution_engine\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mrun_backward\u001b[49m\u001b[43m(\u001b[49m\u001b[43m  \u001b[49m\u001b[38;5;66;43;03m# Calls into the C++ engine to run the backward pass\u001b[39;49;00m\n\u001b[1;32m    198\u001b[0m \u001b[43m    \u001b[49m\u001b[43mtensors\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mgrad_tensors_\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mretain_graph\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mcreate_graph\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43minputs\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m    199\u001b[0m \u001b[43m    \u001b[49m\u001b[43mallow_unreachable\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[38;5;28;43;01mTrue\u001b[39;49;00m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43maccumulate_grad\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[38;5;28;43;01mTrue\u001b[39;49;00m\u001b[43m)\u001b[49m\n",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m: "
     ]
    }
   ],
   "source": [
    "transformer_model = get_model(model_sampler, config, should_train = True)\n",
    "model_path = os.path.join(results_directory, f'bayesian_models_transformer_checkpoint_{model_spec_size}_epochs_'+config['epochs']+'.cpkt')\n",
    "torch.save((transformer_model[2].state_dict(), None), model_path)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Evaluating Transformer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Using cuda device\n",
      "DataLoader.__dict__ {'num_steps': 100, 'fuse_x_y': False, 'get_batch_kwargs': {'batch_size': 256, 'seq_len': 300, 'num_outputs': 1, 'num_features': 8, 'canonical_args': None, 'model': <function <lambda> at 0x7f5c992e1310>}, 'num_features': 8, 'num_outputs': 1}\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "0it [00:00, ?it/s]\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<All keys matched successfully>"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "loaded_epoch = config['epochs']\n",
    "transformer_model = get_model(model_sampler, config, should_train = False)\n",
    "path = os.path.join(results_directory, f'bayesian_models_transformer_checkpoint_{model_spec_size}_epochs_{loaded_epoch}.cpkt')\n",
    "model_state, optimizer_state = torch.load(path)\n",
    "transformer_model[2].load_state_dict(model_state)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "X, y = generate_toy_data(test_model, bptt, device)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2\n",
      "7\n",
      "12\n",
      "17\n",
      "22\n",
      "27\n",
      "32\n",
      "37\n",
      "42\n",
      "47\n",
      "52\n",
      "57\n",
      "62\n",
      "67\n",
      "72\n",
      "77\n",
      "82\n",
      "87\n",
      "92\n"
     ]
    }
   ],
   "source": [
    "results_acc = []\n",
    "results_nll = []\n",
    "transformer_model[2].eval()\n",
    "for training_samples_n in evaluation_points:\n",
    "    print(training_samples_n)\n",
    "    acc, nll, elapsed = eval_transformer(X, y, model=transformer_model[2], training_samples_n=training_samples_n, device=device)\n",
    "    results_acc.append(acc)\n",
    "    results_nll.append(nll)\n",
    "mean = np.array([compute_mean_and_conf_interval(nll)[0] for nll in results_nll])\n",
    "conf = np.array([compute_mean_and_conf_interval(nll)[1] for nll in results_nll])\n",
    "global_results['transformer'] = (None, np.array(evaluation_points), mean, conf)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Plotting results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "../results/timing_big_model_test/results_svi_1_training_samples_2.npy\n",
      "../results/timing_big_model_test/results_svi_1_training_samples_7.npy\n",
      "../results/timing_big_model_test/results_svi_1_training_samples_12.npy\n",
      "../results/timing_big_model_test/results_svi_1_training_samples_17.npy\n",
      "../results/timing_big_model_test/results_svi_1_training_samples_22.npy\n",
      "../results/timing_big_model_test/results_svi_1_training_samples_27.npy\n",
      "../results/timing_big_model_test/results_svi_1_training_samples_32.npy\n",
      "../results/timing_big_model_test/results_svi_1_training_samples_37.npy\n",
      "../results/timing_big_model_test/results_svi_1_training_samples_42.npy\n",
      "../results/timing_big_model_test/results_svi_1_training_samples_47.npy\n",
      "../results/timing_big_model_test/results_svi_1_training_samples_52.npy\n",
      "../results/timing_big_model_test/results_svi_1_training_samples_57.npy\n",
      "../results/timing_big_model_test/results_svi_1_training_samples_62.npy\n",
      "../results/timing_big_model_test/results_svi_1_training_samples_67.npy\n",
      "../results/timing_big_model_test/results_svi_1_training_samples_72.npy\n",
      "../results/timing_big_model_test/results_svi_1_training_samples_77.npy\n",
      "../results/timing_big_model_test/results_svi_1_training_samples_82.npy\n",
      "../results/timing_big_model_test/results_svi_1_training_samples_87.npy\n",
      "../results/timing_big_model_test/results_svi_1_training_samples_92.npy\n",
      "../results/timing_big_model_test/results_mcmc_1_training_samples_2.npy\n",
      "../results/timing_big_model_test/results_mcmc_1_training_samples_7.npy\n",
      "../results/timing_big_model_test/results_mcmc_1_training_samples_12.npy\n",
      "../results/timing_big_model_test/results_mcmc_1_training_samples_17.npy\n",
      "../results/timing_big_model_test/results_mcmc_1_training_samples_22.npy\n",
      "../results/timing_big_model_test/results_mcmc_1_training_samples_27.npy\n",
      "../results/timing_big_model_test/results_mcmc_1_training_samples_32.npy\n",
      "../results/timing_big_model_test/results_mcmc_1_training_samples_37.npy\n",
      "../results/timing_big_model_test/results_mcmc_1_training_samples_42.npy\n",
      "../results/timing_big_model_test/results_mcmc_1_training_samples_47.npy\n",
      "../results/timing_big_model_test/results_mcmc_1_training_samples_52.npy\n",
      "../results/timing_big_model_test/results_mcmc_1_training_samples_57.npy\n",
      "../results/timing_big_model_test/results_mcmc_1_training_samples_62.npy\n",
      "../results/timing_big_model_test/results_mcmc_1_training_samples_67.npy\n",
      "../results/timing_big_model_test/results_mcmc_1_training_samples_72.npy\n",
      "../results/timing_big_model_test/results_mcmc_1_training_samples_77.npy\n",
      "../results/timing_big_model_test/results_mcmc_1_training_samples_82.npy\n",
      "../results/timing_big_model_test/results_mcmc_1_training_samples_87.npy\n",
      "../results/timing_big_model_test/results_mcmc_1_training_samples_92.npy\n"
     ]
    }
   ],
   "source": [
    "files, times, samples, mean, conf = load_results(f'{results_directory}/timing_{model_spec_size}_model_test/results_svi_1_training_{task}', task=task)\n",
    "global_results['svi'] = (times/100, samples, mean/20, conf/30) # original: without /20 /30\n",
    "files, times, samples, mean, conf = load_results(f'{results_directory}/timing_{model_spec_size}_model_test/results_mcmc_1_training_{task}', task=task)\n",
    "global_results['mcmc'] = (times/100, samples,mean/20, conf/30) # original: without /20 /30\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 700x400 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "y_min = min([global_results[k][2].min() for k in global_results])\n",
    "y_max = max([global_results[k][2].max() for k in global_results])\n",
    "\n",
    "fig2 = plt.figure(constrained_layout=True, figsize=(7, 4))\n",
    "axes = plt.axes()\n",
    "axes.set_xlim(0, 100)\n",
    "axes.set_ylim(0,3)\n",
    "import scienceplots\n",
    "with plt.style.context(['ieee','notebook']):\n",
    "    for k in global_results:\n",
    "        plot_with_confidence_intervals(plt, global_results[k][1], global_results[k][2], global_results[k][3], label=k)\n",
    "        #plt.plot(global_results_train_steps[k][1], global_results_train_steps[k][0], label=k)\n",
    "    plt.xlabel('Number of revealed data points(n)',fontsize=14)\n",
    "    plt.ylabel('NLLs',fontsize=14)\n",
    "    plt.legend(loc=\"upper right\")\n",
    "    plt.savefig('../test_curves/compare_other_algorithm/bayesian_svi_mcmc.png')\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "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": "3.9.16"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}