diff --git a/tasks/incentive_classifier/notebooks/BinaryClassifierGoogleColab.ipynb b/tasks/incentive_classifier/notebooks/BinaryClassifierGoogleColab.ipynb
index 114409e8..f6079d70 100644
--- a/tasks/incentive_classifier/notebooks/BinaryClassifierGoogleColab.ipynb
+++ b/tasks/incentive_classifier/notebooks/BinaryClassifierGoogleColab.ipynb
@@ -1,1436 +1,1462 @@
 {
- "cells": [
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {
-    "id": "pqEZnDaKkKzF"
-   },
-   "outputs": [],
-   "source": [
-    "! pip install \\\n",
-    "  scprep\\\n",
-    "  spacy==2.3.2 \\\n",
-    "  sentence_transformers==0.4.0 \\\n",
-    "  phate==1.0.4 && \\\n",
-    "  python -m spacy download es_core_news_lg"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "metadata": {
-    "id": "H41XHYlrkctL"
-   },
-   "source": [
-    "<span style=\"color:red\"><strong>WARNING!</strong></span> Once you installed the packages in the previous cell you must restart your runtime and then import the library and load the model"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {
-    "colab": {
-     "base_uri": "https://localhost:8080/"
-    },
-    "id": "RWngKIdpkPPg",
-    "outputId": "84e989ef-8d85-44bd-8e13-e90aae903469"
-   },
-   "outputs": [
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "Using the GPU\n"
-     ]
-    }
-   ],
-   "source": [
-    "import spacy\n",
-    "if spacy.prefer_gpu():\n",
-    "    print(\"Using the GPU\")\n",
-    "else:\n",
-    "    print(\"Using the CPU\")\n",
-    "es_nlp = spacy.load('es_core_news_lg')"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "metadata": {
-    "id": "55q23OCMkvD5"
-   },
-   "source": [
-    "For development work, in case you want to update the files in your GitHub branch by rerunning the clone, you first have to empty the folder."
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {
-    "id": "Sr1ACGUikx5K"
-   },
-   "outputs": [],
-   "source": [
-    "!rm -rf policy-data-analyzer/"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {
-    "colab": {
-     "base_uri": "https://localhost:8080/"
-    },
-    "id": "FDSRezbdkxyI",
-    "outputId": "7594b505-daf5-4489-e872-89171aaadb8d"
-   },
-   "outputs": [
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "Cloning into 'policy-data-analyzer'...\n",
-      "remote: Enumerating objects: 498, done.\u001b[K\n",
-      "remote: Counting objects: 100% (498/498), done.\u001b[K\n",
-      "remote: Compressing objects: 100% (319/319), done.\u001b[K\n",
-      "remote: Total 2900 (delta 318), reused 337 (delta 178), pack-reused 2402\u001b[K\n",
-      "Receiving objects: 100% (2900/2900), 126.76 MiB | 26.37 MiB/s, done.\n",
-      "Resolving deltas: 100% (1476/1476), done.\n",
-      "Checking out files: 100% (843/843), done.\n"
-     ]
-    }
-   ],
-   "source": [
-    "# Define branch to clone\n",
-    "! branch_name='#50_dfq_sbert_fine_tuning' && \\\n",
-    "  git clone --branch $branch_name https://github.com/wri-dssg/policy-data-analyzer.git"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {
+  "nbformat": 4,
+  "nbformat_minor": 0,
+  "metadata": {
     "colab": {
-     "base_uri": "https://localhost:8080/"
-    },
-    "id": "56PH58KwkxsS",
-    "outputId": "533d36b7-e741-4e38-a3ed-f130f04bb4cd"
-   },
-   "outputs": [
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "Mounted at /content/drive\n"
-     ]
-    }
-   ],
-   "source": [
-    "import pandas as pd\n",
-    "import sys\n",
-    "import os\n",
-    "import csv\n",
-    "from sklearn.model_selection import train_test_split\n",
-    "from sentence_transformers import SentencesDataset, SentenceTransformer, InputExample, losses\n",
-    "from sentence_transformers.evaluation import LabelAccuracyEvaluator\n",
-    "from torch import nn, Tensor\n",
-    "from typing import Iterable, Dict\n",
-    "from torch.utils.data import DataLoader\n",
-    "import math\n",
-    "import time\n",
-    "import cupy as cp\n",
-    "\n",
-    "# os.chdir(\"policy-data-analyzer\") #If you run this cell more than once, comment out this line because you are ready in this folder and you will get an error\n",
-    "from tasks.data_loader.src.utils import *\n",
-    "from tasks.data_augmentation.src.zero_shot_classification.latent_embeddings_classifier import *\n",
-    "from tasks.evaluate_model.src.model_evaluator import *\n",
-    "from tasks.data_visualization.src.plotting import *\n",
-    "\n",
-    "from google.colab import drive\n",
-    "drive.mount('/content/drive')"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "metadata": {
-    "id": "0E6y9u8rmllu"
-   },
-   "source": [
-    "## 1. Fine tune SBERT on Binary Classification"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "metadata": {
-    "id": "uEEBX_uPnQRN"
-   },
-   "source": [
-    "Similar setup for fine tuning multi-class. Maybe we should have this section in the other notebook, and in this one we just load the model? \n"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {
-    "id": "zWb8sqiXkxme"
-   },
-   "outputs": [],
-   "source": [
-    "class SoftmaxClassifier(nn.Module):\n",
-    "    \"\"\"\n",
-    "    This loss adds a softmax classifier on top of the output of the transformer network. \n",
-    "    It takes a sentence embedding and learns a mapping between it and the corresponding category.\n",
-    "    :param model: SentenceTransformer model\n",
-    "    :param sentence_embedding_dimension: Dimension of your sentence embeddings\n",
-    "    :param num_labels: Number of different labels\n",
-    "    \"\"\"\n",
-    "    def __init__(self,\n",
-    "                 model: SentenceTransformer,\n",
-    "                 sentence_embedding_dimension: int,\n",
-    "                 num_labels: int):\n",
-    "        super(SoftmaxClassifier, self).__init__()\n",
-    "        self.model = model\n",
-    "        self.num_labels = num_labels\n",
-    "        self.classifier = nn.Linear(sentence_embedding_dimension, num_labels)\n",
-    "\n",
-    "    def forward(self, sentence_features: Iterable[Dict[str, Tensor]], labels: Tensor):\n",
-    "        # Get batch sentence embeddings\n",
-    "        features = self.model(sentence_features[0])['sentence_embedding']\n",
-    "        \n",
-    "        # Get batch loss\n",
-    "        output = self.classifier(features)\n",
-    "        loss_fct = nn.CrossEntropyLoss()\n",
-    "\n",
-    "        if labels is not None:\n",
-    "            loss = loss_fct(output, labels.view(-1))\n",
-    "            return loss\n",
-    "        else:\n",
-    "            return features, output"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "metadata": {
-    "id": "plcg1u8kqCCR"
-   },
-   "source": [
-    "### Load data\n",
-    "For now, EXP9 will be binary classification"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {
-    "id": "liOnrCm-kCVg"
-   },
-   "outputs": [],
-   "source": [
-    "rater = \"Rater3\" # TODO: Change accordingly to what is the dataset you want to analyze\n",
-    "Experiment = \"EXP9\" \n",
-    "\n",
-    "# This first one is the one used by David and Daniel\n",
-    "path = \"/content/drive/MyDrive/WRI-LatinAmerica-Talent/Cristina_Policy_Files/Tagged_sentence_lists/datasets/\"\n",
-    "# This one is the one used by Jordi\n",
-    "# path = \"/content/drive/MyDrive/Official Folder of WRI Latin America Project/WRI-LatinAmerica-Talent/Cristina_Policy_Files/Tagged_sentence_lists/datasets/\"\n",
-    "\n",
-    "model_names = ['stsb-xlm-r-multilingual', 'paraphrase-xlm-r-multilingual-v1']#, 'quora-distilbert-multilingual''distiluse-base-multilingual-cased-v2', "
-   ]
+      "name": "BinaryClassifierGoogleColab.ipynb",
+      "provenance": [],
+      "collapsed_sections": [],
+      "machine_shape": "hm"
+    },
+    "kernelspec": {
+      "name": "python3",
+      "display_name": "Python 3"
+    },
+    "accelerator": "GPU"
   },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {
-    "colab": {
-     "base_uri": "https://localhost:8080/"
+  "cells": [
+    {
+      "cell_type": "code",
+      "metadata": {
+        "id": "pqEZnDaKkKzF"
+      },
+      "source": [
+        "! pip install \\\n",
+        "  scprep\\\n",
+        "  spacy==2.3.2 \\\n",
+        "  sentence_transformers==0.4.0 \\\n",
+        "  phate==1.0.4 && \\\n",
+        "  python -m spacy download es_core_news_lg"
+      ],
+      "execution_count": null,
+      "outputs": []
     },
-    "id": "CnN_0CAVkJRt",
-    "outputId": "4d822392-3d5a-4956-a61a-20ffc9dcf5ee"
-   },
-   "outputs": [
     {
-     "data": {
-      "text/plain": [
-       "['not_Incentive', 'Incentive']"
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "H41XHYlrkctL"
+      },
+      "source": [
+        "<span style=\"color:red\"><strong>WARNING!</strong></span> Once you installed the packages in the previous cell you must restart your runtime and then import the library and load the model"
       ]
-     },
-     "execution_count": 18,
-     "metadata": {
-      "tags": []
-     },
-     "output_type": "execute_result"
-    }
-   ],
-   "source": [
-    "filename = \"dataset_\" + rater + \"_incentive_sentences.csv\"\n",
-    "file = path + filename\n",
-    "with open(file, newline='') as f:\n",
-    "    reader = csv.reader(f)\n",
-    "    all_sents = list(reader)[0]\n",
-    "\n",
-    "filename = \"dataset_\" + rater + \"_incentive_labels.csv\"\n",
-    "file = path + filename\n",
-    "with open(file, newline='') as f:\n",
-    "    reader = csv.reader(f)\n",
-    "    all_labels = list(reader)[0]\n",
-    "\n",
-    "filename = \"testset_\" + rater + \"_incentive_sentences.csv\"\n",
-    "file = path + filename\n",
-    "with open(file, newline='') as f:\n",
-    "    reader = csv.reader(f)\n",
-    "    test_sents = list(reader)[0]\n",
-    "\n",
-    "filename = \"testset_\" + rater + \"_incentive_labels.csv\"\n",
-    "file = path + filename\n",
-    "with open(file, newline='') as f:\n",
-    "    reader = csv.reader(f)\n",
-    "    test_labels = list(reader)[0]\n",
-    "\n",
-    "label_names = unique_labels(all_labels)\n",
-    "numeric_labels = labels2numeric(all_labels, label_names)\n",
-    "label_names"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {
-    "colab": {
-     "base_uri": "https://localhost:8080/"
-    },
-    "id": "jmj6mDkxo42F",
-    "outputId": "35cf0641-0ce3-4e1e-94dd-8668679bfa84"
-   },
-   "outputs": [
-    {
-     "data": {
-      "text/plain": [
-       "('El pago de los recibos se hara en todo caso correlativamente, no siendo admisible el pago de uno de ellos dejando pendiente el anterior o anteriores',\n",
-       " 'not_Incentive',\n",
-       " 'Facilitar el acceso a mercados, servicios financieros, programas y proyectos.',\n",
-       " 'Incentive')"
+    },
+    {
+      "cell_type": "code",
+      "metadata": {
+        "colab": {
+          "base_uri": "https://localhost:8080/"
+        },
+        "id": "RWngKIdpkPPg",
+        "outputId": "d48bc45c-616b-4c5a-c432-c84630978360"
+      },
+      "source": [
+        "import spacy\n",
+        "if spacy.prefer_gpu():\n",
+        "    print(\"Using the GPU\")\n",
+        "else:\n",
+        "    print(\"Using the CPU\")\n",
+        "es_nlp = spacy.load('es_core_news_lg')"
+      ],
+      "execution_count": 1,
+      "outputs": [
+        {
+          "output_type": "stream",
+          "text": [
+            "Using the GPU\n"
+          ],
+          "name": "stdout"
+        }
       ]
-     },
-     "execution_count": 19,
-     "metadata": {
-      "tags": []
-     },
-     "output_type": "execute_result"
-    }
-   ],
-   "source": [
-    "all_sents[2], all_labels[2], test_sents[2], test_labels[2]"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {
-    "colab": {
-     "base_uri": "https://localhost:8080/",
-     "height": 387
-    },
-    "id": "ij2Z9Hi4tWT4",
-    "outputId": "79ddc730-885f-4f0e-9c1d-cb25768c0cf3"
-   },
-   "outputs": [
-    {
-     "data": {
-      "image/png": "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\n",
-      "text/plain": [
-       "<Figure size 432x288 with 1 Axes>"
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "55q23OCMkvD5"
+      },
+      "source": [
+        "For development work, in case you want to update the files in your GitHub branch by rerunning the clone, you first have to empty the folder."
       ]
-     },
-     "metadata": {
-      "needs_background": "light",
-      "tags": []
-     },
-     "output_type": "display_data"
-    },
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "Label counts:\n",
-      "{'not_Incentive': 0.2532786885245902, 'Incentive': 0.7467213114754099}\n"
-     ]
-    }
-   ],
-   "source": [
-    "plot_data_distribution(numeric_labels, label_names)"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "metadata": {
-    "id": "bE911nx4uzfq"
-   },
-   "source": [
-    "As we can see, we have 3/4 of the data as Incentives and 1/4 as Non-incentives. Hopefully our model will be able to differentiate!"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {
-    "id": "7iwJ0mc8uzNH"
-   },
-   "outputs": [],
-   "source": []
-  },
-  {
-   "cell_type": "markdown",
-   "metadata": {
-    "id": "Pm9OVVBXqFVj"
-   },
-   "source": [
-    "### Replicate Fine Tuning from Multi Class\n",
-    "\n",
-    "However, for time purposes, we will use the minimum number of parameters necessary so we can get to the binary classification setup part! When actually fine tuning, we should explore more hyperparameters. Things to change when running full:\n",
-    "- Add more test/validation data percentages in `all_test_perc`\n",
-    "- Include more models in `model_names`\n",
-    "- Comment out `num_epochs=10`, replace parameter `epochs` in `model.fit()` to be 2, uncomment `max_num_epochs` and uncomment for loop that iterates through epochs. \n",
-    "- Comment out the parameter `output_path` in `model.fit()`\n",
-    "- **IMPORTANT:** Since we don't store the models during normal training, we should create the directory in which the confusion matrix and t-SNE plots should be written to"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {
-    "id": "FrAaQLo6pI4m"
-   },
-   "outputs": [],
-   "source": [
-    "# Train test split stratified\n",
-    "all_test_perc = [0.2]\n",
-    "model_names = ['paraphrase-xlm-r-multilingual-v1']#, 'stsb-xlm-r-multilingual', 'quora-distilbert-multilingual', 'distiluse-base-multilingual-cased-v2', \n",
-    "\n",
-    "# Output setup\n",
-    "output = {}\n",
-    "\n",
-    "for test_perc in all_test_perc:\n",
-    "  output[f\"test_perc={test_perc}\"] = {}\n",
-    "  X_train, X_test, y_train, y_test = train_test_split(all_sents, all_labels, test_size=test_perc, stratify=all_labels, random_state=69420)\n",
-    "\n",
-    "  # Load data samples into batches\n",
-    "  train_batch_size = 16\n",
-    "  label2int = dict(zip(label_names, range(len(label_names))))\n",
-    "  train_samples = []\n",
-    "  for sent, label in zip(X_train, y_train):\n",
-    "      label_id = label2int[label]\n",
-    "      train_samples.append(InputExample(texts=[sent], label=label_id))\n",
-    "\n",
-    "  # Configure the dev set evaluator - still need to test whether this works\n",
-    "  dev_samples = []\n",
-    "  for sent, label in zip(X_test, y_test):\n",
-    "      label_id = label2int[label]\n",
-    "      dev_samples.append(InputExample(texts=[sent], label=label_id))\n",
-    "  \n",
-    "  for model_name in model_names:\n",
-    "    # Setup\n",
-    "    model_preds = []\n",
-    "    model_scores = []\n",
-    "    output[f\"test_perc={test_perc}\"][model_name] = []\n",
-    "    \n",
-    "    # Train set config\n",
-    "    model = SentenceTransformer(model_name)\n",
-    "    train_dataset = SentencesDataset(train_samples, model=model)\n",
-    "    train_dataloader = DataLoader(train_dataset, shuffle=True, batch_size=train_batch_size)\n",
-    "    \n",
-    "    # Define the way the loss is computed\n",
-    "    classifier = SoftmaxClassifier(model=model, sentence_embedding_dimension=model.get_sentence_embedding_dimension(), num_labels=len(label2int))\n",
-    "    \n",
-    "    # Dev set config\n",
-    "    dev_dataset = SentencesDataset(dev_samples, model=model)\n",
-    "    dev_dataloader = DataLoader(dev_dataset, shuffle=True, batch_size=train_batch_size)\n",
-    "    dev_evaluator = LabelAccuracyEvaluator(dataloader=dev_dataloader, softmax_model=classifier, name='lae-dev')\n",
-    "\n",
-    "    # Configure the training\n",
-    "    # max_num_epochs = 10\n",
-    "    num_epochs = 10 # For now we will make the model train straight up to 10 epochs\n",
-    "    \n",
-    "    # for num_epochs in range(4, max_num_epochs + 2, 2):\n",
-    "    print(\"Num epochs:\", num_epochs)\n",
-    "    \n",
-    "    warmup_steps = math.ceil(len(train_dataset) * num_epochs / train_batch_size * 0.1)  # 10% of train data for warm-up\n",
-    "    model_deets = f\"model={model_name}_test-perc={test_perc}_n-epoch={num_epochs}\"\n",
-    "  \n",
-    "    model_save_path = f\"/content/drive/MyDrive/WRI-LatinAmerica-Talent/Modeling/FineTuningExperiments/{Experiment}/BinaryFineTuning_{model_deets}\"\n",
-    "    # model_save_path = f\"/content/drive/MyDrive/Official Folder of WRI Latin America Project/WRI-LatinAmerica-Talent/Modeling/FineTuningExperiments/{Experiment}/FineTuning_{model_deets}\"\n",
-    "    \n",
-    "\n",
-    "    # Train the model\n",
-    "    start = time.time()\n",
-    "\n",
-    "    # WARNING: epochs should be changed to 2 when testing multiple epochs... or maybe we should have another variable for it\n",
-    "    model.fit(train_objectives=[(train_dataloader, classifier)],\n",
-    "              evaluator=dev_evaluator,\n",
-    "              epochs=num_epochs, # We always tune on an extra epoch to see the performance gain\n",
-    "              evaluation_steps=1000,\n",
-    "              warmup_steps=warmup_steps,\n",
-    "              output_path=model_save_path\n",
-    "              )\n",
-    "    \n",
-    "    end = time.time()\n",
-    "    hours, rem = divmod(end-start, 3600)\n",
-    "    minutes, seconds = divmod(rem, 60)\n",
-    "    print(\"Time taken for fine-tuning:\", \"{:0>2}:{:0>2}:{:05.2f}\".format(int(hours),int(minutes),seconds))\n",
-    "    \n",
-    "    ### Classify sentences\n",
-    "    # Projection matrix Z low-dim projection\n",
-    "    print(\"Classifying sentences...\")\n",
-    "    proj_matrix = cp.asnumpy(calc_proj_matrix(test_sents, 50, es_nlp, model, 0.01))\n",
-    "    all_sent_embs = encode_all_sents(test_sents, model, proj_matrix)\n",
-    "    all_label_embs = encode_labels(label_names, model, proj_matrix)\n",
-    "    visualize_embeddings_2D(np.vstack(all_sent_embs), test_labels, tsne_perplexity=50, store_name=f\"{model_save_path}/{model_deets}\")\n",
-    "    model_preds, model_scores = calc_all_cos_similarity(all_sent_embs, all_label_embs, label_names)\n",
-    "    \n",
-    "    ### Evaluate the model\n",
-    "    numeric_preds = labels2numeric(model_preds, label_names)\n",
-    "    numeric_test_labels = labels2numeric(test_labels, label_names)\n",
-    "    evaluator = ModelEvaluator(label_names, y_true=numeric_test_labels, y_pred=numeric_preds)\n",
-    "    \n",
-    "    output[f\"test_perc={test_perc}\"][model_name].append({\"num_epochs\": num_epochs, \"avg_f1\": evaluator.avg_f1.tolist()})\n",
-    "    \n",
-    "    evaluator.plot_confusion_matrix(color_map='Blues', exp_name=f\"{model_save_path}/{model_deets}\")"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {
-    "id": "VrIcnp1Wse60"
-   },
-   "outputs": [],
-   "source": [
-    "new_json = {}\n",
-    "\n",
-    "for key in output.keys():\n",
-    "  new_json[key] = {}\n",
-    "  for subkey in output[key].keys():\n",
-    "    new_json[key][subkey] = []\n",
-    "    for element in output[key][subkey]:\n",
-    "      el_copy = {\"avg_f1\": element[\"avg_f1\"], \"num_epochs\": element[\"num_epochs\"]}\n",
-    "      new_json[key][subkey].append(el_copy)"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {
-    "id": "aPlLyprt0ILI"
-   },
-   "outputs": [],
-   "source": [
-    "import json\n",
-    "output_path = f\"/content/drive/MyDrive/WRI-LatinAmerica-Talent/Modeling/FineTuningExperiments/{Experiment}/\"\n",
-    "# output_path = f\"/content/drive/MyDrive/Official Folder of WRI Latin America Project/WRI-LatinAmerica-Talent/Modeling/FineTuningExperiments/{Experiment}/\"\n",
-    "with open(f\"{output_path}{Experiment}_FineTuningResults.json\", \"w\") as f:\n",
-    "  json.dump(new_json, f)"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "metadata": {
-    "id": "wLx3D9nrtrDE"
-   },
-   "source": [
-    "## 2. Load fine-tuned model and use embeddings to train a binary classifier"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "metadata": {
-    "id": "TepO8Pr3wsOX"
-   },
-   "source": [
-    "### Load model"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {
-    "id": "i_0lKmiGtw0N"
-   },
-   "outputs": [],
-   "source": [
-    "model_name = \"paraphrase-xlm-r-multilingual-v1\"\n",
-    "test_perc = 0.2\n",
-    "num_epochs = 10\n",
-    "model_deets = f\"model={model_name}_test-perc={test_perc}_n-epoch={num_epochs}\"\n",
-    "experiment = \"EXP9\"\n",
-    "saved_model_path = f\"/content/drive/MyDrive/WRI-LatinAmerica-Talent/Modeling/FineTuningExperiments/{experiment}/BinaryFineTuning_{model_deets}\"\n",
-    "# saved_model_path = f\"/content/drive/MyDrive/Official Folder of WRI Latin America Project/WRI-LatinAmerica-Talent/Modeling/FineTuningExperiments/{Experiment}/FineTuning_{model_deets}\""
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {
-    "id": "dHYaPvP5wZc_"
-   },
-   "outputs": [],
-   "source": [
-    "bin_model = SentenceTransformer(saved_model_path)"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "metadata": {
-    "id": "iY4NIIWbwu_M"
-   },
-   "source": [
-    "### Encode Sentences"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "metadata": {
-    "id": "t98Vpa-51t2K"
-   },
-   "source": [
-    "First, we will check how good are the fine tuned embeddings without the projection matrix addition"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {
-    "colab": {
-     "base_uri": "https://localhost:8080/"
-    },
-    "id": "7Qv5doEyw7_r",
-    "outputId": "16e98bc1-6994-4996-e382-069ff6d3fe84"
-   },
-   "outputs": [
-    {
-     "name": "stderr",
-     "output_type": "stream",
-     "text": [
-      "100%|██████████| 306/306 [00:05<00:00, 54.29it/s]\n"
-     ]
-    }
-   ],
-   "source": [
-    "all_sent_embs = encode_all_sents(test_sents, bin_model)"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {
-    "colab": {
-     "base_uri": "https://localhost:8080/",
-     "height": 412
-    },
-    "id": "XcEIw-sb1sFm",
-    "outputId": "0a6d1fad-c4c0-4d54-a905-c27fc005f7d7"
-   },
-   "outputs": [
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "[t-SNE] Computing 151 nearest neighbors...\n",
-      "[t-SNE] Indexed 306 samples in 0.033s...\n",
-      "[t-SNE] Computed neighbors for 306 samples in 0.217s...\n",
-      "[t-SNE] Computed conditional probabilities for sample 306 / 306\n",
-      "[t-SNE] Mean sigma: 2.491708\n",
-      "[t-SNE] KL divergence after 250 iterations with early exaggeration: 59.511703\n",
-      "[t-SNE] KL divergence after 1000 iterations: 0.449039\n"
-     ]
-    },
-    {
-     "data": {
-      "image/png": "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\n",
-      "text/plain": [
-       "<Figure size 1728x288 with 2 Axes>"
+    },
+    {
+      "cell_type": "code",
+      "metadata": {
+        "id": "Sr1ACGUikx5K"
+      },
+      "source": [
+        "!rm -rf policy-data-analyzer/"
+      ],
+      "execution_count": 2,
+      "outputs": []
+    },
+    {
+      "cell_type": "code",
+      "metadata": {
+        "colab": {
+          "base_uri": "https://localhost:8080/"
+        },
+        "id": "FDSRezbdkxyI",
+        "outputId": "294fdc80-81e2-41d7-9739-713fd5e523c8"
+      },
+      "source": [
+        "# Define branch to clone\n",
+        "! branch_name='#50_dfq_sbert_fine_tuning' && \\\n",
+        "  git clone --branch $branch_name https://github.com/wri-dssg/policy-data-analyzer.git"
+      ],
+      "execution_count": 3,
+      "outputs": [
+        {
+          "output_type": "stream",
+          "text": [
+            "Cloning into 'policy-data-analyzer'...\n",
+            "remote: Enumerating objects: 75, done.\u001b[K\n",
+            "remote: Counting objects: 100% (75/75), done.\u001b[K\n",
+            "remote: Compressing objects: 100% (52/52), done.\u001b[K\n",
+            "remote: Total 2994 (delta 47), reused 48 (delta 23), pack-reused 2919\u001b[K\n",
+            "Receiving objects: 100% (2994/2994), 129.28 MiB | 10.94 MiB/s, done.\n",
+            "Resolving deltas: 100% (1594/1594), done.\n",
+            "Checking out files: 100% (843/843), done.\n"
+          ],
+          "name": "stdout"
+        }
       ]
-     },
-     "metadata": {
-      "tags": []
-     },
-     "output_type": "display_data"
-    }
-   ],
-   "source": [
-    "visualize_embeddings_2D(np.vstack(all_sent_embs), test_labels, tsne_perplexity=50)"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "metadata": {
-    "id": "5m3G_4eF2Tw4"
-   },
-   "source": [
-    "Ok, doesn't look that bad, but not perfect either... The incentives are scattered too much in the space, and the line between non-incentives and incentives is not clearly defined. ***For now, it doesn't matter - but we should experiment more with fine tuning.***.\n",
-    "\n",
-    "Now, let's check whether the projection matrix helps:"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {
-    "colab": {
-     "base_uri": "https://localhost:8080/"
-    },
-    "id": "auWe9D9S10mD",
-    "outputId": "9795754c-7182-4e75-b33b-e04c31dec83d"
-   },
-   "outputs": [
-    {
-     "name": "stderr",
-     "output_type": "stream",
-     "text": [
-      "100%|██████████| 306/306 [00:07<00:00, 40.82it/s]\n"
-     ]
-    }
-   ],
-   "source": [
-    "proj_matrix = cp.asnumpy(calc_proj_matrix(all_sents, 50, es_nlp, bin_model, 0.01))\n",
-    "all_sent_embs = encode_all_sents(test_sents, bin_model, proj_matrix)"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {
-    "colab": {
-     "base_uri": "https://localhost:8080/",
-     "height": 412
-    },
-    "id": "55hNSqtS19JE",
-    "outputId": "3a17d090-265b-40e0-de35-f2287f11efea"
-   },
-   "outputs": [
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "[t-SNE] Computing 151 nearest neighbors...\n",
-      "[t-SNE] Indexed 306 samples in 0.006s...\n",
-      "[t-SNE] Computed neighbors for 306 samples in 0.083s...\n",
-      "[t-SNE] Computed conditional probabilities for sample 306 / 306\n",
-      "[t-SNE] Mean sigma: 20.201386\n",
-      "[t-SNE] KL divergence after 250 iterations with early exaggeration: 69.918625\n",
-      "[t-SNE] KL divergence after 1000 iterations: 0.734087\n"
-     ]
-    },
-    {
-     "data": {
-      "image/png": "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\n",
-      "text/plain": [
-       "<Figure size 1728x288 with 2 Axes>"
+    },
+    {
+      "cell_type": "code",
+      "metadata": {
+        "colab": {
+          "base_uri": "https://localhost:8080/"
+        },
+        "id": "56PH58KwkxsS",
+        "outputId": "73fc15d3-ffb5-4ec9-ff8b-fe1b35e5dadb"
+      },
+      "source": [
+        "import pandas as pd\n",
+        "import sys\n",
+        "import os\n",
+        "import csv\n",
+        "from sklearn.model_selection import train_test_split\n",
+        "from sentence_transformers import SentencesDataset, SentenceTransformer, InputExample, losses\n",
+        "from sentence_transformers.evaluation import LabelAccuracyEvaluator\n",
+        "from torch import nn, Tensor\n",
+        "from typing import Iterable, Dict\n",
+        "from torch.utils.data import DataLoader\n",
+        "import math\n",
+        "import time\n",
+        "import cupy as cp\n",
+        "\n",
+        "os.chdir(\"policy-data-analyzer\") #If you run this cell more than once, comment out this line because you are ready in this folder and you will get an error\n",
+        "from tasks.data_loader.src.utils import *\n",
+        "from tasks.data_augmentation.src.zero_shot_classification.latent_embeddings_classifier import *\n",
+        "from tasks.evaluate_model.src.model_evaluator import *\n",
+        "from tasks.data_visualization.src.plotting import *\n",
+        "\n",
+        "from google.colab import drive\n",
+        "drive.mount('/content/drive')"
+      ],
+      "execution_count": 4,
+      "outputs": [
+        {
+          "output_type": "stream",
+          "text": [
+            "Drive already mounted at /content/drive; to attempt to forcibly remount, call drive.mount(\"/content/drive\", force_remount=True).\n"
+          ],
+          "name": "stdout"
+        }
       ]
-     },
-     "metadata": {
-      "tags": []
-     },
-     "output_type": "display_data"
-    }
-   ],
-   "source": [
-    "visualize_embeddings_2D(np.vstack(all_sent_embs), test_labels, tsne_perplexity=50)"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "metadata": {
-    "id": "BqqfluJR2-0B"
-   },
-   "source": [
-    "Actually, the projection matrix makes things worse. ***Let's NOT use it for now!!!!!***"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {
-    "colab": {
-     "base_uri": "https://localhost:8080/"
-    },
-    "id": "pLPUbG282_SO",
-    "outputId": "57745c51-8804-4dbf-ff60-9d5b475b8440"
-   },
-   "outputs": [
-    {
-     "name": "stderr",
-     "output_type": "stream",
-     "text": [
-      "100%|██████████| 1220/1220 [00:22<00:00, 54.73it/s]\n"
-     ]
-    }
-   ],
-   "source": [
-    "# Simple embeddings, no projection matrix added\n",
-    "all_sent_embs = encode_all_sents(all_sents, bin_model)"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 68,
-   "metadata": {
-    "colab": {
-     "base_uri": "https://localhost:8080/"
-    },
-    "id": "3lvfAZQjAHvS",
-    "outputId": "41e1c935-6308-406a-bef8-014a4d941e5d"
-   },
-   "outputs": [
-    {
-     "name": "stderr",
-     "output_type": "stream",
-     "text": [
-      "100%|██████████| 306/306 [00:05<00:00, 54.31it/s]\n"
-     ]
-    }
-   ],
-   "source": [
-    "all_test_embs = encode_all_sents(test_sents, bin_model)"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "metadata": {
-    "id": "qyLt6Pl84UHc"
-   },
-   "source": [
-    "### Train classifiers"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 120,
-   "metadata": {
-    "id": "vCR_wG18E7aC"
-   },
-   "outputs": [],
-   "source": [
-    "from sklearn.model_selection import cross_val_score\n",
-    "from sklearn.metrics import classification_report"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "metadata": {
-    "id": "rQyQIFVEBZsk"
-   },
-   "source": [
-    "1. Let's start with Random Forests!"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 64,
-   "metadata": {
-    "id": "YW021S1J_JH9"
-   },
-   "outputs": [],
-   "source": [
-    "from sklearn.ensemble import RandomForestClassifier"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 135,
-   "metadata": {
-    "id": "dbQ3_grn_I82"
-   },
-   "outputs": [],
-   "source": [
-    "clf = RandomForestClassifier(n_estimators=100, max_depth=3, random_state=69420)"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 136,
-   "metadata": {
-    "colab": {
-     "base_uri": "https://localhost:8080/"
-    },
-    "id": "fAM7nFv3_IxT",
-    "outputId": "54c22d6a-7303-42d7-cdd3-819dad2633b8"
-   },
-   "outputs": [
-    {
-     "data": {
-      "text/plain": [
-       "RandomForestClassifier(bootstrap=True, ccp_alpha=0.0, class_weight=None,\n",
-       "                       criterion='gini', max_depth=3, max_features='auto',\n",
-       "                       max_leaf_nodes=None, max_samples=None,\n",
-       "                       min_impurity_decrease=0.0, min_impurity_split=None,\n",
-       "                       min_samples_leaf=1, min_samples_split=2,\n",
-       "                       min_weight_fraction_leaf=0.0, n_estimators=100,\n",
-       "                       n_jobs=None, oob_score=False, random_state=69420,\n",
-       "                       verbose=0, warm_start=False)"
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "0E6y9u8rmllu"
+      },
+      "source": [
+        "## 1. Fine tune SBERT on Binary Classification"
       ]
-     },
-     "execution_count": 136,
-     "metadata": {
-      "tags": []
-     },
-     "output_type": "execute_result"
-    }
-   ],
-   "source": [
-    "clf.fit(np.vstack(all_sent_embs), all_labels)"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 137,
-   "metadata": {
-    "id": "0MsZbft8_6_1"
-   },
-   "outputs": [],
-   "source": [
-    "clf_preds = [clf.predict(sent_emb)[0] for sent_emb in all_test_embs]"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 138,
-   "metadata": {
-    "colab": {
-     "base_uri": "https://localhost:8080/"
-    },
-    "id": "2uLnsvS6GVVk",
-    "outputId": "5346cb46-8f87-4a18-83d8-be9391c26027"
-   },
-   "outputs": [
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "               precision    recall  f1-score   support\n",
-      "\n",
-      "    Incentive       0.89      0.93      0.91       228\n",
-      "not_Incentive       0.76      0.68      0.72        78\n",
-      "\n",
-      "     accuracy                           0.86       306\n",
-      "    macro avg       0.83      0.80      0.81       306\n",
-      " weighted avg       0.86      0.86      0.86       306\n",
-      "\n"
-     ]
-    }
-   ],
-   "source": [
-    " print(classification_report(test_labels, clf_preds))"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 139,
-   "metadata": {
-    "id": "Y_k1CWeJATgy"
-   },
-   "outputs": [],
-   "source": [
-    "numeric_preds = labels2numeric(clf_preds, label_names)\n",
-    "numeric_test_labels = labels2numeric(test_labels, label_names)\n",
-    "evaluator = ModelEvaluator(label_names, y_true=numeric_test_labels, y_pred=numeric_preds)"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 140,
-   "metadata": {
-    "colab": {
-     "base_uri": "https://localhost:8080/",
-     "height": 455
-    },
-    "id": "jHI9hf6EAy_N",
-    "outputId": "1544a62f-3b4c-4aaf-b98b-82d9d3044783"
-   },
-   "outputs": [
-    {
-     "data": {
-      "image/png": "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\n",
-      "text/plain": [
-       "<Figure size 576x432 with 2 Axes>"
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "uEEBX_uPnQRN"
+      },
+      "source": [
+        "Similar setup for fine tuning multi-class. Maybe we should have this section in the other notebook, and in this one we just load the model? \n"
       ]
-     },
-     "metadata": {
-      "tags": []
-     },
-     "output_type": "display_data"
-    }
-   ],
-   "source": [
-    "evaluator.plot_confusion_matrix(color_map='Blues')"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "metadata": {
-    "id": "yXdaChWFBrWg"
-   },
-   "source": [
-    "Honestly, without Grid Search and 5-fold Cross Validation, these are not bad results... We should add those though!"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "metadata": {
-    "id": "Y3DoYgdaDHJ6"
-   },
-   "source": [
-    "2. Now, we're gonna try Support Vector Machines"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 99,
-   "metadata": {
-    "id": "lcNVNbZVDBX_"
-   },
-   "outputs": [],
-   "source": [
-    "from sklearn import svm"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 129,
-   "metadata": {
-    "id": "vD4-V3KpAy3k"
-   },
-   "outputs": [],
-   "source": [
-    "clf = svm.SVC(gamma=0.001, C=100.)"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 130,
-   "metadata": {
-    "colab": {
-     "base_uri": "https://localhost:8080/"
-    },
-    "id": "oeqA23VGAyoK",
-    "outputId": "3d905ff1-b4e8-4d2b-8887-5a6eb5e497e1"
-   },
-   "outputs": [
-    {
-     "data": {
-      "text/plain": [
-       "SVC(C=100.0, break_ties=False, cache_size=200, class_weight=None, coef0=0.0,\n",
-       "    decision_function_shape='ovr', degree=3, gamma=0.001, kernel='rbf',\n",
-       "    max_iter=-1, probability=False, random_state=None, shrinking=True,\n",
-       "    tol=0.001, verbose=False)"
+    },
+    {
+      "cell_type": "code",
+      "metadata": {
+        "id": "zWb8sqiXkxme"
+      },
+      "source": [
+        "class SoftmaxClassifier(nn.Module):\n",
+        "    \"\"\"\n",
+        "    This loss adds a softmax classifier on top of the output of the transformer network. \n",
+        "    It takes a sentence embedding and learns a mapping between it and the corresponding category.\n",
+        "    :param model: SentenceTransformer model\n",
+        "    :param sentence_embedding_dimension: Dimension of your sentence embeddings\n",
+        "    :param num_labels: Number of different labels\n",
+        "    \"\"\"\n",
+        "    def __init__(self,\n",
+        "                 model: SentenceTransformer,\n",
+        "                 sentence_embedding_dimension: int,\n",
+        "                 num_labels: int):\n",
+        "        super(SoftmaxClassifier, self).__init__()\n",
+        "        self.model = model\n",
+        "        self.num_labels = num_labels\n",
+        "        self.classifier = nn.Linear(sentence_embedding_dimension, num_labels)\n",
+        "\n",
+        "    def forward(self, sentence_features: Iterable[Dict[str, Tensor]], labels: Tensor):\n",
+        "        # Get batch sentence embeddings\n",
+        "        features = self.model(sentence_features[0])['sentence_embedding']\n",
+        "        \n",
+        "        # Get batch loss\n",
+        "        output = self.classifier(features)\n",
+        "        loss_fct = nn.CrossEntropyLoss()\n",
+        "\n",
+        "        if labels is not None:\n",
+        "            loss = loss_fct(output, labels.view(-1))\n",
+        "            return loss\n",
+        "        else:\n",
+        "            return features, output"
+      ],
+      "execution_count": 5,
+      "outputs": []
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "plcg1u8kqCCR"
+      },
+      "source": [
+        "### Load data\n",
+        "\n",
+        "- EXP1 = Binary classification fine tuning using the projection matrix\n",
+        "- EXP2 = Binary classification fine tuning ***without*** the projection matrix"
       ]
-     },
-     "execution_count": 130,
-     "metadata": {
-      "tags": []
-     },
-     "output_type": "execute_result"
-    }
-   ],
-   "source": [
-    "clf.fit(np.vstack(all_sent_embs), all_labels)"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 131,
-   "metadata": {
-    "id": "tyytG01oEC4G"
-   },
-   "outputs": [],
-   "source": [
-    "clf_preds = [clf.predict(sent_emb)[0] for sent_emb in all_test_embs]"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 132,
-   "metadata": {
-    "id": "SEneBVYFD3DW"
-   },
-   "outputs": [],
-   "source": [
-    "numeric_preds = labels2numeric(clf_preds, label_names)\n",
-    "numeric_test_labels = labels2numeric(test_labels, label_names)\n",
-    "evaluator = ModelEvaluator(label_names, y_true=numeric_test_labels, y_pred=numeric_preds)"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 133,
-   "metadata": {
-    "colab": {
-     "base_uri": "https://localhost:8080/"
-    },
-    "id": "cscCU733D664",
-    "outputId": "51dd0385-c7ae-44f6-f090-936b4161136a"
-   },
-   "outputs": [
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "               precision    recall  f1-score   support\n",
-      "\n",
-      "    Incentive       0.91      0.92      0.92       228\n",
-      "not_Incentive       0.76      0.74      0.75        78\n",
-      "\n",
-      "     accuracy                           0.88       306\n",
-      "    macro avg       0.84      0.83      0.84       306\n",
-      " weighted avg       0.87      0.88      0.88       306\n",
-      "\n"
-     ]
-    }
-   ],
-   "source": [
-    "print(classification_report(test_labels, clf_preds))"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 134,
-   "metadata": {
-    "colab": {
-     "base_uri": "https://localhost:8080/",
-     "height": 455
-    },
-    "id": "oxYzT51aD60j",
-    "outputId": "c7af7030-d09a-43a0-c868-5ac36af22b7a"
-   },
-   "outputs": [
-    {
-     "data": {
-      "image/png": "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\n",
-      "text/plain": [
-       "<Figure size 576x432 with 2 Axes>"
+    },
+    {
+      "cell_type": "code",
+      "metadata": {
+        "id": "liOnrCm-kCVg"
+      },
+      "source": [
+        "rater = \"Rater3\" # TODO: Change accordingly to what is the dataset you want to analyze\n",
+        "Experiment = \"EXP2\" \n",
+        "\n",
+        "# This first one is the one used by David and Daniel\n",
+        "base_path = \"/content/drive/MyDrive/WRI-LatinAmerica-Talent\"\n",
+        "\n",
+        "# This one is the one used by Jordi\n",
+        "# base_path = \"/content/drive/MyDrive/Official Folder of WRI Latin America Project/WRI-LatinAmerica-Talent\"\n",
+        "\n",
+        "data_path = f\"{base_path}/Cristina_Policy_Files/Tagged_sentence_lists/datasets/\"\n",
+        "results_save_path = f\"{base_path}/Modeling/BinaryClassificationExperiments/{Experiment}\"\n",
+        "\n",
+        "if not os.path.exists(results_save_path):\n",
+        "    os.makedirs(results_save_path)"
+      ],
+      "execution_count": 6,
+      "outputs": []
+    },
+    {
+      "cell_type": "code",
+      "metadata": {
+        "colab": {
+          "base_uri": "https://localhost:8080/"
+        },
+        "id": "CnN_0CAVkJRt",
+        "outputId": "f2ff82a7-9a12-43cb-dd0e-21fbd0dc33c5"
+      },
+      "source": [
+        "filename = \"dataset_\" + rater + \"_incentive_sentences.csv\"\n",
+        "file = data_path + filename\n",
+        "with open(file, newline='') as f:\n",
+        "    reader = csv.reader(f)\n",
+        "    all_sents = list(reader)[0]\n",
+        "\n",
+        "filename = \"dataset_\" + rater + \"_incentive_labels.csv\"\n",
+        "file = data_path + filename\n",
+        "with open(file, newline='') as f:\n",
+        "    reader = csv.reader(f)\n",
+        "    all_labels = list(reader)[0]\n",
+        "\n",
+        "filename = \"testset_\" + rater + \"_incentive_sentences.csv\"\n",
+        "file = data_path + filename\n",
+        "with open(file, newline='') as f:\n",
+        "    reader = csv.reader(f)\n",
+        "    test_sents = list(reader)[0]\n",
+        "\n",
+        "filename = \"testset_\" + rater + \"_incentive_labels.csv\"\n",
+        "file = data_path + filename\n",
+        "with open(file, newline='') as f:\n",
+        "    reader = csv.reader(f)\n",
+        "    test_labels = list(reader)[0]\n",
+        "\n",
+        "label_names = unique_labels(all_labels)\n",
+        "numeric_labels = labels2numeric(all_labels, label_names)\n",
+        "label_names"
+      ],
+      "execution_count": 7,
+      "outputs": [
+        {
+          "output_type": "execute_result",
+          "data": {
+            "text/plain": [
+              "['not_Incentive', 'Incentive']"
+            ]
+          },
+          "metadata": {
+            "tags": []
+          },
+          "execution_count": 7
+        }
       ]
-     },
-     "metadata": {
-      "tags": []
-     },
-     "output_type": "display_data"
-    }
-   ],
-   "source": [
-    "evaluator.plot_confusion_matrix(color_map='Blues')"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "metadata": {
-    "id": "H-ki8dU5EUaq"
-   },
-   "source": [
-    "Ok, so SVMs are slightly better than Random Forests at ***differentiating*** text! There's a 1-2% decrease in performance for the incentive class, but a 6% gain in non-incentives. If these results remain when doing cross validation and grid search, then I'd recommend going for the SVMs.\n",
-    "\n",
-    "\n",
-    "**Next steps:**\n",
-    "- Add Grid Search Cross Validation from sklearn"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {
-    "id": "9y34yQNqD6TO"
-   },
-   "outputs": [],
-   "source": []
-  },
-  {
-   "cell_type": "markdown",
-   "metadata": {
-    "id": "mbfAvxu0-_Qi"
-   },
-   "source": [
-    "### What about... Beto?\n",
-    "I downloaded the weights and placed them in the folder below: "
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 142,
-   "metadata": {
-    "id": "mxk0ZHUFHnkk"
-   },
-   "outputs": [],
-   "source": [
-    "beto_path = f\"/content/drive/MyDrive/WRI-LatinAmerica-Talent/Modeling/BETO/pytorch/\""
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 145,
-   "metadata": {
-    "id": "7lHUcXtY3PKP"
-   },
-   "outputs": [],
-   "source": [
-    "from transformers import AutoTokenizer, AutoModelForSequenceClassification\n",
-    "from transformers import BertTokenizer, BertForSequenceClassification"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "metadata": {
-    "id": "TI9IF0UmKYZE"
-   },
-   "source": [
-    "***IMPORTANT:*** I was not able to figure out a way of using the fine tuning results from the models above so I'm gonna use BETO out of the box, for both encoding/classification and see how it goes.\n",
-    "\n",
-    "The following cells are a demo of how the model should be put to use - once you understand it, feel free to skip this part!"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 146,
-   "metadata": {
-    "colab": {
-     "base_uri": "https://localhost:8080/"
-    },
-    "id": "kmJ_mpC37ICQ",
-    "outputId": "d3d7a3f6-8717-4e50-d554-18779d622b1e"
-   },
-   "outputs": [
-    {
-     "name": "stderr",
-     "output_type": "stream",
-     "text": [
-      "Some weights of the model checkpoint at /content/drive/MyDrive/WRI-LatinAmerica-Talent/Modeling/BETO/pytorch/ were not used when initializing BertForSequenceClassification: ['cls.predictions.bias', 'cls.predictions.transform.dense.weight', 'cls.predictions.transform.dense.bias', 'cls.predictions.transform.LayerNorm.weight', 'cls.predictions.transform.LayerNorm.bias', 'cls.predictions.decoder.weight', 'cls.seq_relationship.weight', 'cls.seq_relationship.bias']\n",
-      "- This IS expected if you are initializing BertForSequenceClassification from the checkpoint of a model trained on another task or with another architecture (e.g. initializing a BertForSequenceClassification model from a BertForPreTraining model).\n",
-      "- This IS NOT expected if you are initializing BertForSequenceClassification from the checkpoint of a model that you expect to be exactly identical (initializing a BertForSequenceClassification model from a BertForSequenceClassification model).\n",
-      "Some weights of BertForSequenceClassification were not initialized from the model checkpoint at /content/drive/MyDrive/WRI-LatinAmerica-Talent/Modeling/BETO/pytorch/ and are newly initialized: ['classifier.weight', 'classifier.bias']\n",
-      "You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.\n"
-     ]
-    }
-   ],
-   "source": [
-    "tokenizer = BertTokenizer.from_pretrained(beto_path)\n",
-    "model = BertForSequenceClassification.from_pretrained(beto_path)"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 148,
-   "metadata": {
-    "id": "EGRL66pBIzuy"
-   },
-   "outputs": [],
-   "source": [
-    "classes = [\"no es parafrasis\", \"es parafrasis\"]\n",
-    "\n",
-    "sequence_0 = \"La compañia Hugging esta basada en Nueva York\"\n",
-    "sequence_1 = \"Las manzanas son malas para la salud\"\n",
-    "sequence_2 = \"La sede principal de Hugging esta en Manhattan\"\n",
-    "\n",
-    "paraphrase = tokenizer(sequence_0, sequence_2, return_tensors=\"pt\")\n",
-    "not_paraphrase = tokenizer(sequence_0, sequence_1, return_tensors=\"pt\")\n",
-    "\n",
-    "paraphrase_classification_logits = model(**paraphrase).logits\n",
-    "not_paraphrase_classification_logits = model(**not_paraphrase).logits\n",
-    "\n",
-    "paraphrase_results = torch.softmax(paraphrase_classification_logits, dim=1).tolist()[0]\n",
-    "not_paraphrase_results = torch.softmax(not_paraphrase_classification_logits, dim=1).tolist()[0]"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 156,
-   "metadata": {
-    "colab": {
-     "base_uri": "https://localhost:8080/"
-    },
-    "id": "C-Pkqe-vI3ld",
-    "outputId": "5d208439-c5ae-4da4-b6a1-e748c3a2f839"
-   },
-   "outputs": [
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      ">>> Deberia de ser parafrasis:\n",
-      "no es parafrasis: 41%\n",
-      "es parafrasis: 59%\n",
-      "\n",
-      ">>> NO deberia de ser parafrasis:\n",
-      "no es parafrasis: 61%\n",
-      "es parafrasis: 39%\n"
-     ]
-    }
-   ],
-   "source": [
-    "print(\">>> Deberia de ser parafrasis:\")\n",
-    "for i in range(len(classes)):\n",
-    "  print(f\"{classes[i]}: {int(round(paraphrase_results[i] * 100))}%\")\n",
-    "\n",
-    "print(\"\\n>>> NO deberia de ser parafrasis:\")\n",
-    "for i in range(len(classes)):\n",
-    "  print(f\"{classes[i]}: {int(round(not_paraphrase_results[i] * 100))}%\")"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "metadata": {
-    "id": "0dRr9sJQKezT"
-   },
-   "source": [
-    "OK! Now it's time to apply it to our data. We will try it out with our test set, just to have a fair comparison"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 159,
-   "metadata": {
-    "colab": {
-     "base_uri": "https://localhost:8080/",
-     "height": 362
-    },
-    "id": "ybHMwJE-JyvG",
-    "outputId": "66e60822-56a7-45ff-c01a-96e9c54ded9b"
-   },
-   "outputs": [
-    {
-     "ename": "RuntimeError",
-     "evalue": "ignored",
-     "output_type": "error",
-     "traceback": [
-      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
-      "\u001b[0;31mRuntimeError\u001b[0m                              Traceback (most recent call last)",
-      "\u001b[0;32m<ipython-input-159-848fd0460984>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[1;32m      1\u001b[0m \u001b[0mtokenized_sents\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mtokenizer\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtest_sents\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mpadding\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mTrue\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mreturn_tensors\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m\"pt\"\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 2\u001b[0;31m \u001b[0mclf_logits\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mmodel\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m**\u001b[0m\u001b[0mtokenized_sents\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mlogits\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      3\u001b[0m \u001b[0mclf_results\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mtorch\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msoftmax\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mclf_logits\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdim\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtolist\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
-      "\u001b[0;32m/usr/local/lib/python3.6/dist-packages/torch/nn/modules/module.py\u001b[0m in \u001b[0;36m_call_impl\u001b[0;34m(self, *input, **kwargs)\u001b[0m\n\u001b[1;32m    725\u001b[0m             \u001b[0mresult\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_slow_forward\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0minput\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    726\u001b[0m         \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 727\u001b[0;31m             \u001b[0mresult\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mforward\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0minput\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    728\u001b[0m         for hook in itertools.chain(\n\u001b[1;32m    729\u001b[0m                 \u001b[0m_global_forward_hooks\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mvalues\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
-      "\u001b[0;32m/usr/local/lib/python3.6/dist-packages/transformers/models/bert/modeling_bert.py\u001b[0m in \u001b[0;36mforward\u001b[0;34m(self, input_ids, attention_mask, token_type_ids, position_ids, head_mask, inputs_embeds, labels, output_attentions, output_hidden_states, return_dict)\u001b[0m\n\u001b[1;32m   1495\u001b[0m             \u001b[0moutput_attentions\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0moutput_attentions\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1496\u001b[0m             \u001b[0moutput_hidden_states\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0moutput_hidden_states\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 1497\u001b[0;31m             \u001b[0mreturn_dict\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mreturn_dict\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   1498\u001b[0m         )\n\u001b[1;32m   1499\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
-      "\u001b[0;32m/usr/local/lib/python3.6/dist-packages/torch/nn/modules/module.py\u001b[0m in \u001b[0;36m_call_impl\u001b[0;34m(self, *input, **kwargs)\u001b[0m\n\u001b[1;32m    725\u001b[0m             \u001b[0mresult\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_slow_forward\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0minput\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    726\u001b[0m         \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 727\u001b[0;31m             \u001b[0mresult\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mforward\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0minput\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    728\u001b[0m         for hook in itertools.chain(\n\u001b[1;32m    729\u001b[0m                 \u001b[0m_global_forward_hooks\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mvalues\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
-      "\u001b[0;32m/usr/local/lib/python3.6/dist-packages/transformers/models/bert/modeling_bert.py\u001b[0m in \u001b[0;36mforward\u001b[0;34m(self, input_ids, attention_mask, token_type_ids, position_ids, head_mask, inputs_embeds, encoder_hidden_states, encoder_attention_mask, past_key_values, use_cache, output_attentions, output_hidden_states, return_dict)\u001b[0m\n\u001b[1;32m    966\u001b[0m             \u001b[0moutput_attentions\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0moutput_attentions\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    967\u001b[0m             \u001b[0moutput_hidden_states\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0moutput_hidden_states\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 968\u001b[0;31m             \u001b[0mreturn_dict\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mreturn_dict\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    969\u001b[0m         )\n\u001b[1;32m    970\u001b[0m         \u001b[0msequence_output\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mencoder_outputs\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
-      "\u001b[0;32m/usr/local/lib/python3.6/dist-packages/torch/nn/modules/module.py\u001b[0m in \u001b[0;36m_call_impl\u001b[0;34m(self, *input, **kwargs)\u001b[0m\n\u001b[1;32m    725\u001b[0m             \u001b[0mresult\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_slow_forward\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0minput\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    726\u001b[0m         \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 727\u001b[0;31m             \u001b[0mresult\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mforward\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0minput\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    728\u001b[0m         for hook in itertools.chain(\n\u001b[1;32m    729\u001b[0m                 \u001b[0m_global_forward_hooks\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mvalues\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
-      "\u001b[0;32m/usr/local/lib/python3.6/dist-packages/transformers/models/bert/modeling_bert.py\u001b[0m in \u001b[0;36mforward\u001b[0;34m(self, hidden_states, attention_mask, head_mask, encoder_hidden_states, encoder_attention_mask, past_key_values, use_cache, output_attentions, output_hidden_states, return_dict)\u001b[0m\n\u001b[1;32m    564\u001b[0m                     \u001b[0mencoder_attention_mask\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    565\u001b[0m                     \u001b[0mpast_key_value\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 566\u001b[0;31m                     \u001b[0moutput_attentions\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    567\u001b[0m                 )\n\u001b[1;32m    568\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
-      "\u001b[0;32m/usr/local/lib/python3.6/dist-packages/torch/nn/modules/module.py\u001b[0m in \u001b[0;36m_call_impl\u001b[0;34m(self, *input, **kwargs)\u001b[0m\n\u001b[1;32m    725\u001b[0m             \u001b[0mresult\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_slow_forward\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0minput\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    726\u001b[0m         \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 727\u001b[0;31m             \u001b[0mresult\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mforward\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0minput\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    728\u001b[0m         for hook in itertools.chain(\n\u001b[1;32m    729\u001b[0m                 \u001b[0m_global_forward_hooks\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mvalues\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
-      "\u001b[0;32m/usr/local/lib/python3.6/dist-packages/transformers/models/bert/modeling_bert.py\u001b[0m in \u001b[0;36mforward\u001b[0;34m(self, hidden_states, attention_mask, head_mask, encoder_hidden_states, encoder_attention_mask, past_key_value, output_attentions)\u001b[0m\n\u001b[1;32m    494\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    495\u001b[0m         layer_output = apply_chunking_to_forward(\n\u001b[0;32m--> 496\u001b[0;31m             \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfeed_forward_chunk\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mchunk_size_feed_forward\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mseq_len_dim\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mattention_output\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    497\u001b[0m         )\n\u001b[1;32m    498\u001b[0m         \u001b[0moutputs\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mlayer_output\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m+\u001b[0m \u001b[0moutputs\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
-      "\u001b[0;32m/usr/local/lib/python3.6/dist-packages/transformers/modeling_utils.py\u001b[0m in \u001b[0;36mapply_chunking_to_forward\u001b[0;34m(forward_fn, chunk_size, chunk_dim, *input_tensors)\u001b[0m\n\u001b[1;32m   1785\u001b[0m         \u001b[0;32mreturn\u001b[0m \u001b[0mtorch\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcat\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0moutput_chunks\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdim\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mchunk_dim\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1786\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 1787\u001b[0;31m     \u001b[0;32mreturn\u001b[0m \u001b[0mforward_fn\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0minput_tensors\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
-      "\u001b[0;32m/usr/local/lib/python3.6/dist-packages/transformers/models/bert/modeling_bert.py\u001b[0m in \u001b[0;36mfeed_forward_chunk\u001b[0;34m(self, attention_output)\u001b[0m\n\u001b[1;32m    505\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    506\u001b[0m     \u001b[0;32mdef\u001b[0m \u001b[0mfeed_forward_chunk\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mattention_output\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 507\u001b[0;31m         \u001b[0mintermediate_output\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mintermediate\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mattention_output\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    508\u001b[0m         \u001b[0mlayer_output\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0moutput\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mintermediate_output\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mattention_output\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    509\u001b[0m         \u001b[0;32mreturn\u001b[0m \u001b[0mlayer_output\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
-      "\u001b[0;32m/usr/local/lib/python3.6/dist-packages/torch/nn/modules/module.py\u001b[0m in \u001b[0;36m_call_impl\u001b[0;34m(self, *input, **kwargs)\u001b[0m\n\u001b[1;32m    725\u001b[0m             \u001b[0mresult\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_slow_forward\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0minput\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    726\u001b[0m         \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 727\u001b[0;31m             \u001b[0mresult\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mforward\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0minput\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    728\u001b[0m         for hook in itertools.chain(\n\u001b[1;32m    729\u001b[0m                 \u001b[0m_global_forward_hooks\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mvalues\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
-      "\u001b[0;32m/usr/local/lib/python3.6/dist-packages/transformers/models/bert/modeling_bert.py\u001b[0m in \u001b[0;36mforward\u001b[0;34m(self, hidden_states)\u001b[0m\n\u001b[1;32m    409\u001b[0m     \u001b[0;32mdef\u001b[0m \u001b[0mforward\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mhidden_states\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    410\u001b[0m         \u001b[0mhidden_states\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdense\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mhidden_states\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 411\u001b[0;31m         \u001b[0mhidden_states\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mintermediate_act_fn\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mhidden_states\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    412\u001b[0m         \u001b[0;32mreturn\u001b[0m \u001b[0mhidden_states\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    413\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
-      "\u001b[0;32m/usr/local/lib/python3.6/dist-packages/torch/nn/functional.py\u001b[0m in \u001b[0;36mgelu\u001b[0;34m(input)\u001b[0m\n\u001b[1;32m   1381\u001b[0m         \u001b[0;32mif\u001b[0m \u001b[0mtype\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0minput\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0mTensor\u001b[0m \u001b[0;32mand\u001b[0m \u001b[0mhas_torch_function\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0minput\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1382\u001b[0m             \u001b[0;32mreturn\u001b[0m \u001b[0mhandle_torch_function\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mgelu\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0minput\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0minput\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 1383\u001b[0;31m     \u001b[0;32mreturn\u001b[0m \u001b[0mtorch\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_C\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_nn\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mgelu\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0minput\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   1384\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1385\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
-      "\u001b[0;31mRuntimeError\u001b[0m: CUDA out of memory. Tried to allocate 768.00 MiB (GPU 0; 14.73 GiB total capacity; 11.33 GiB already allocated; 413.88 MiB free; 12.41 GiB reserved in total by PyTorch)"
-     ]
-    }
-   ],
-   "source": [
-    "tokenized_sents = tokenizer(test_sents, padding=True, return_tensors=\"pt\")\n",
-    "clf_logits = model(**tokenized_sents).logits\n",
-    "clf_results = torch.softmax(clf_logits, dim=1).tolist()[0]"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {
-    "id": "wELlB7iUJ1T3"
-   },
-   "outputs": [],
-   "source": [
-    "# This stores the index of the highest score - in other words, our label\n",
-    "clf_preds = [np.argmax(logits) for logits in clf_results]"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {
-    "id": "GnrG0AwcM-G5"
-   },
-   "outputs": [],
-   "source": [
-    "print(classification_report(test_labels, clf_preds))"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 161,
-   "metadata": {
-    "colab": {
-     "base_uri": "https://localhost:8080/"
     },
-    "id": "qdV91B4IMZdz",
-    "outputId": "c6903dcb-c967-428f-a157-42717317da82"
-   },
-   "outputs": [
     {
-     "data": {
-      "text/plain": [
-       "1"
+      "cell_type": "code",
+      "metadata": {
+        "colab": {
+          "base_uri": "https://localhost:8080/"
+        },
+        "id": "jmj6mDkxo42F",
+        "outputId": "ab072c4d-f317-46f6-f4f7-e49dce446f45"
+      },
+      "source": [
+        "all_sents[2], all_labels[2], test_sents[2], test_labels[2]"
+      ],
+      "execution_count": 8,
+      "outputs": [
+        {
+          "output_type": "execute_result",
+          "data": {
+            "text/plain": [
+              "('El pago de los recibos se hara en todo caso correlativamente, no siendo admisible el pago de uno de ellos dejando pendiente el anterior o anteriores',\n",
+              " 'not_Incentive',\n",
+              " 'Facilitar el acceso a mercados, servicios financieros, programas y proyectos.',\n",
+              " 'Incentive')"
+            ]
+          },
+          "metadata": {
+            "tags": []
+          },
+          "execution_count": 8
+        }
+      ]
+    },
+    {
+      "cell_type": "code",
+      "metadata": {
+        "colab": {
+          "base_uri": "https://localhost:8080/",
+          "height": 387
+        },
+        "id": "ij2Z9Hi4tWT4",
+        "outputId": "66e403a1-ec8b-4775-9846-a0efea3d985e"
+      },
+      "source": [
+        "plot_data_distribution(numeric_labels, label_names)"
+      ],
+      "execution_count": 9,
+      "outputs": [
+        {
+          "output_type": "display_data",
+          "data": {
+            "image/png": "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\n",
+            "text/plain": [
+              "<Figure size 432x288 with 1 Axes>"
+            ]
+          },
+          "metadata": {
+            "tags": [],
+            "needs_background": "light"
+          }
+        },
+        {
+          "output_type": "stream",
+          "text": [
+            "Label counts:\n",
+            "{'not_Incentive': 0.2532786885245902, 'Incentive': 0.7467213114754099}\n"
+          ],
+          "name": "stdout"
+        }
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "bE911nx4uzfq"
+      },
+      "source": [
+        "As we can see, we have 3/4 of the data as Incentives and 1/4 as Non-incentives. Hopefully our model will be able to differentiate!"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "metadata": {
+        "id": "7iwJ0mc8uzNH"
+      },
+      "source": [
+        ""
+      ],
+      "execution_count": 9,
+      "outputs": []
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "Pm9OVVBXqFVj"
+      },
+      "source": [
+        "### Replicate Fine Tuning from Multi Class\n",
+        "\n",
+        "However, for time purposes, we will use the minimum number of parameters necessary so we can get to the binary classification setup part! When actually fine tuning, we should explore more hyperparameters. Things to change when running full:\n",
+        "- Add more test/validation data percentages in `all_test_perc`\n",
+        "- Include more models in `model_names`\n",
+        "- Comment out `num_epochs=10`, replace parameter `epochs` in `model.fit()` to be 2, uncomment `max_num_epochs` and uncomment for loop that iterates through epochs. \n",
+        "- Comment out the parameter `output_path` in `model.fit()`\n",
+        "- **IMPORTANT:** Since we don't store the model output path"
       ]
-     },
-     "execution_count": 161,
-     "metadata": {
-      "tags": []
-     },
-     "output_type": "execute_result"
+    },
+    {
+      "cell_type": "code",
+      "metadata": {
+        "id": "a5vcnVTrCPa8"
+      },
+      "source": [
+        "import json"
+      ],
+      "execution_count": 10,
+      "outputs": []
+    },
+    {
+      "cell_type": "code",
+      "metadata": {
+        "id": "irVDlABFCRSp"
+      },
+      "source": [
+        "output_path = f\"{base_path}/Modeling/BinaryClassificationExperiments/{Experiment}/\"\n",
+        "# output_path = f\"/content/drive/MyDrive/Official Folder of WRI Latin America Project/WRI-LatinAmerica-Talent/Modeling/FineTuningExperiments/{Experiment}/\""
+      ],
+      "execution_count": 11,
+      "outputs": []
+    },
+    {
+      "cell_type": "code",
+      "metadata": {
+        "id": "FrAaQLo6pI4m"
+      },
+      "source": [
+        "# Configure the training\n",
+        "all_test_perc = [0.15, 0.2, 0.25]\n",
+        "model_names = ['paraphrase-xlm-r-multilingual-v1', 'stsb-xlm-r-multilingual', 'quora-distilbert-multilingual', 'distiluse-base-multilingual-cased-v2']\n",
+        "\n",
+        "# If you want to train for a set number of epochs instead of a range, set all these numbers to be equal\n",
+        "start_epochs = 4\n",
+        "epochs_increment = 2\n",
+        "max_num_epochs = 12\n",
+        "\n",
+        "\n",
+        "# Output setup - we will update the json as the fine tuning process goes so every result is stored immediately\n",
+        "with open(f\"{output_path}{Experiment}_FineTuningResults.json\", \"w\") as fw:\n",
+        "  json.dump({}, fw)\n",
+        "\n",
+        "for test_perc in all_test_perc:\n",
+        "  with open(f\"{output_path}{Experiment}_FineTuningResults.json\", \"r\") as fr:\n",
+        "    output = json.load(fr)\n",
+        "\n",
+        "  output[f\"test_perc={test_perc}\"] = {}\n",
+        "  X_train, X_test, y_train, y_test = train_test_split(all_sents, all_labels, test_size=test_perc, stratify=all_labels, random_state=69420)\n",
+        "\n",
+        "  # Load data samples into batches\n",
+        "  train_batch_size = 16\n",
+        "  label2int = dict(zip(label_names, range(len(label_names))))\n",
+        "  train_samples = []\n",
+        "  for sent, label in zip(X_train, y_train):\n",
+        "      label_id = label2int[label]\n",
+        "      train_samples.append(InputExample(texts=[sent], label=label_id))\n",
+        "\n",
+        "  # Configure the dev set evaluator - still need to test whether this works\n",
+        "  dev_samples = []\n",
+        "  for sent, label in zip(X_test, y_test):\n",
+        "      label_id = label2int[label]\n",
+        "      dev_samples.append(InputExample(texts=[sent], label=label_id))\n",
+        "  \n",
+        "  for model_name in model_names:\n",
+        "    # Setup\n",
+        "    model_preds = []\n",
+        "    model_scores = []\n",
+        "    output[f\"test_perc={test_perc}\"][model_name] = []\n",
+        "    \n",
+        "    # Train set config\n",
+        "    model = SentenceTransformer(model_name)\n",
+        "    train_dataset = SentencesDataset(train_samples, model=model)\n",
+        "    train_dataloader = DataLoader(train_dataset, shuffle=True, batch_size=train_batch_size)\n",
+        "    \n",
+        "    # Define the way the loss is computed\n",
+        "    classifier = SoftmaxClassifier(model=model, sentence_embedding_dimension=model.get_sentence_embedding_dimension(), num_labels=len(label2int))\n",
+        "    \n",
+        "    # Dev set config\n",
+        "    dev_dataset = SentencesDataset(dev_samples, model=model)\n",
+        "    dev_dataloader = DataLoader(dev_dataset, shuffle=True, batch_size=train_batch_size)\n",
+        "    dev_evaluator = LabelAccuracyEvaluator(dataloader=dev_dataloader, softmax_model=classifier, name='lae-dev')\n",
+        "    \n",
+        "    for num_epochs in range(start_epochs, max_num_epochs + 2, epochs_increment):\n",
+        "      print(\"Num epochs:\", num_epochs)\n",
+        "      \n",
+        "      warmup_steps = math.ceil(len(train_dataset) * num_epochs / train_batch_size * 0.1)  # 10% of train data for warm-up\n",
+        "      model_deets = f\"model={model_name}_test-perc={test_perc}_n-epoch={num_epochs}\"\n",
+        "    \n",
+        "      # Train the model\n",
+        "      start = time.time()\n",
+        "      \n",
+        "      if num_epochs == start_epochs:\n",
+        "        model.fit(train_objectives=[(train_dataloader, classifier)],\n",
+        "                  evaluator=dev_evaluator,\n",
+        "                  epochs=start_epochs, \n",
+        "                  evaluation_steps=1000,\n",
+        "                  warmup_steps=warmup_steps,\n",
+        "                  output_path=f\"{results_save_path}/BinaryCls_{model_deets}\"\n",
+        "                  )\n",
+        "      else:\n",
+        "        model.fit(train_objectives=[(train_dataloader, classifier)],\n",
+        "                  evaluator=dev_evaluator,\n",
+        "                  epochs=epochs_increment, # We always tune on an extra epoch to see the performance gain\n",
+        "                  evaluation_steps=1000,\n",
+        "                  warmup_steps=warmup_steps,\n",
+        "                  output_path=f\"{results_save_path}/BinaryCls_{model_deets}\"\n",
+        "                  )\n",
+        "        \n",
+        "      end = time.time()\n",
+        "      hours, rem = divmod(end-start, 3600)\n",
+        "      minutes, seconds = divmod(rem, 60)\n",
+        "      print(\"Time taken for fine-tuning:\", \"{:0>2}:{:0>2}:{:05.2f}\".format(int(hours),int(minutes),seconds))\n",
+        "      \n",
+        "      ### Classify sentences\n",
+        "      # Projection matrix Z low-dim projection\n",
+        "      print(\"Classifying sentences...\")\n",
+        "      proj_matrix = cp.asnumpy(calc_proj_matrix(test_sents, 50, es_nlp, model, 0.01))\n",
+        "      all_sent_embs = encode_all_sents(test_sents, model, proj_matrix)\n",
+        "      all_label_embs = encode_labels(label_names, model, proj_matrix)\n",
+        "      visualize_embeddings_2D(np.vstack(all_sent_embs), test_labels, tsne_perplexity=50, store_name=f\"{results_save_path}/BinaryCls_{model_deets}/{model_deets}\")\n",
+        "      model_preds, model_scores = calc_all_cos_similarity(all_sent_embs, all_label_embs, label_names)\n",
+        "      \n",
+        "      ### Evaluate the model\n",
+        "      numeric_preds = labels2numeric(model_preds, label_names)\n",
+        "      numeric_test_labels = labels2numeric(test_labels, label_names)\n",
+        "      evaluator = ModelEvaluator(label_names, y_true=numeric_test_labels, y_pred=numeric_preds)\n",
+        "      \n",
+        "      output[f\"test_perc={test_perc}\"][model_name].append({\"num_epochs\": num_epochs, \"avg_f1\": evaluator.avg_f1.tolist()})\n",
+        "      with open(f\"{output_path}{Experiment}_FineTuningResults.json\", \"w\") as fw:\n",
+        "        json.dump(output, fw)\n",
+        "\n",
+        "      evaluator.plot_confusion_matrix(color_map='Blues', exp_name=f\"{results_save_path}/BinaryCls_{model_deets}/{model_deets}\")"
+      ],
+      "execution_count": null,
+      "outputs": []
+    },
+    {
+      "cell_type": "code",
+      "metadata": {
+        "id": "VrIcnp1Wse60"
+      },
+      "source": [
+        "new_json = {}\n",
+        "\n",
+        "for key in output.keys():\n",
+        "  new_json[key] = {}\n",
+        "  for subkey in output[key].keys():\n",
+        "    new_json[key][subkey] = []\n",
+        "    for element in output[key][subkey]:\n",
+        "      el_copy = {\"avg_f1\": element[\"avg_f1\"], \"num_epochs\": element[\"num_epochs\"]}\n",
+        "      new_json[key][subkey].append(el_copy)"
+      ],
+      "execution_count": null,
+      "outputs": []
+    },
+    {
+      "cell_type": "code",
+      "metadata": {
+        "id": "aPlLyprt0ILI"
+      },
+      "source": [
+        "import json\n",
+        "output_path = f\"/content/drive/MyDrive/WRI-LatinAmerica-Talent/Modeling/FineTuningExperiments/{Experiment}/\"\n",
+        "# output_path = f\"/content/drive/MyDrive/Official Folder of WRI Latin America Project/WRI-LatinAmerica-Talent/Modeling/FineTuningExperiments/{Experiment}/\"\n",
+        "with open(f\"{output_path}{Experiment}_FineTuningResults.json\", \"w\") as f:\n",
+        "  json.dump(new_json, f)"
+      ],
+      "execution_count": null,
+      "outputs": []
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "wLx3D9nrtrDE"
+      },
+      "source": [
+        "## 2. Load fine-tuned model and use embeddings to train a binary classifier"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "TepO8Pr3wsOX"
+      },
+      "source": [
+        "### Load model"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "metadata": {
+        "id": "i_0lKmiGtw0N"
+      },
+      "source": [
+        "model_name = \"paraphrase-xlm-r-multilingual-v1\"\n",
+        "test_perc = 0.2\n",
+        "num_epochs = 10\n",
+        "model_deets = f\"model={model_name}_test-perc={test_perc}_n-epoch={num_epochs}\"\n",
+        "experiment = \"EXP9\"\n",
+        "saved_model_path = f\"{base_path}/Modeling/FineTuningExperiments/BinaryFineTuning_{experiment}\"\n",
+        "# saved_model_path = f\"/content/drive/MyDrive/Official Folder of WRI Latin America Project/WRI-LatinAmerica-Talent/Modeling/FineTuningExperiments/{Experiment}/FineTuning_{model_deets}\""
+      ],
+      "execution_count": null,
+      "outputs": []
+    },
+    {
+      "cell_type": "code",
+      "metadata": {
+        "id": "dHYaPvP5wZc_"
+      },
+      "source": [
+        "bin_model = SentenceTransformer(saved_model_path)"
+      ],
+      "execution_count": null,
+      "outputs": []
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "iY4NIIWbwu_M"
+      },
+      "source": [
+        "### Encode Sentences"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "t98Vpa-51t2K"
+      },
+      "source": [
+        "First, we will check how good are the fine tuned embeddings without the projection matrix addition"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "metadata": {
+        "colab": {
+          "base_uri": "https://localhost:8080/"
+        },
+        "id": "7Qv5doEyw7_r",
+        "outputId": "16e98bc1-6994-4996-e382-069ff6d3fe84"
+      },
+      "source": [
+        "all_sent_embs = encode_all_sents(test_sents, bin_model)"
+      ],
+      "execution_count": null,
+      "outputs": [
+        {
+          "output_type": "stream",
+          "text": [
+            "100%|██████████| 306/306 [00:05<00:00, 54.29it/s]\n"
+          ],
+          "name": "stderr"
+        }
+      ]
+    },
+    {
+      "cell_type": "code",
+      "metadata": {
+        "colab": {
+          "base_uri": "https://localhost:8080/",
+          "height": 412
+        },
+        "id": "XcEIw-sb1sFm",
+        "outputId": "0a6d1fad-c4c0-4d54-a905-c27fc005f7d7"
+      },
+      "source": [
+        "visualize_embeddings_2D(np.vstack(all_sent_embs), test_labels, tsne_perplexity=50)"
+      ],
+      "execution_count": null,
+      "outputs": [
+        {
+          "output_type": "stream",
+          "text": [
+            "[t-SNE] Computing 151 nearest neighbors...\n",
+            "[t-SNE] Indexed 306 samples in 0.033s...\n",
+            "[t-SNE] Computed neighbors for 306 samples in 0.217s...\n",
+            "[t-SNE] Computed conditional probabilities for sample 306 / 306\n",
+            "[t-SNE] Mean sigma: 2.491708\n",
+            "[t-SNE] KL divergence after 250 iterations with early exaggeration: 59.511703\n",
+            "[t-SNE] KL divergence after 1000 iterations: 0.449039\n"
+          ],
+          "name": "stdout"
+        },
+        {
+          "output_type": "display_data",
+          "data": {
+            "image/png": "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\n",
+            "text/plain": [
+              "<Figure size 1728x288 with 2 Axes>"
+            ]
+          },
+          "metadata": {
+            "tags": []
+          }
+        }
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "5m3G_4eF2Tw4"
+      },
+      "source": [
+        "Ok, doesn't look that bad, but not perfect either... The incentives are scattered too much in the space, and the line between non-incentives and incentives is not clearly defined. ***For now, it doesn't matter - but we should experiment more with fine tuning.***.\n",
+        "\n",
+        "Now, let's check whether the projection matrix helps:"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "metadata": {
+        "colab": {
+          "base_uri": "https://localhost:8080/"
+        },
+        "id": "auWe9D9S10mD",
+        "outputId": "9795754c-7182-4e75-b33b-e04c31dec83d"
+      },
+      "source": [
+        "proj_matrix = cp.asnumpy(calc_proj_matrix(all_sents, 50, es_nlp, bin_model, 0.01))\n",
+        "all_sent_embs = encode_all_sents(test_sents, bin_model, proj_matrix)"
+      ],
+      "execution_count": null,
+      "outputs": [
+        {
+          "output_type": "stream",
+          "text": [
+            "100%|██████████| 306/306 [00:07<00:00, 40.82it/s]\n"
+          ],
+          "name": "stderr"
+        }
+      ]
+    },
+    {
+      "cell_type": "code",
+      "metadata": {
+        "colab": {
+          "base_uri": "https://localhost:8080/",
+          "height": 412
+        },
+        "id": "55hNSqtS19JE",
+        "outputId": "3a17d090-265b-40e0-de35-f2287f11efea"
+      },
+      "source": [
+        "visualize_embeddings_2D(np.vstack(all_sent_embs), test_labels, tsne_perplexity=50)"
+      ],
+      "execution_count": null,
+      "outputs": [
+        {
+          "output_type": "stream",
+          "text": [
+            "[t-SNE] Computing 151 nearest neighbors...\n",
+            "[t-SNE] Indexed 306 samples in 0.006s...\n",
+            "[t-SNE] Computed neighbors for 306 samples in 0.083s...\n",
+            "[t-SNE] Computed conditional probabilities for sample 306 / 306\n",
+            "[t-SNE] Mean sigma: 20.201386\n",
+            "[t-SNE] KL divergence after 250 iterations with early exaggeration: 69.918625\n",
+            "[t-SNE] KL divergence after 1000 iterations: 0.734087\n"
+          ],
+          "name": "stdout"
+        },
+        {
+          "output_type": "display_data",
+          "data": {
+            "image/png": "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\n",
+            "text/plain": [
+              "<Figure size 1728x288 with 2 Axes>"
+            ]
+          },
+          "metadata": {
+            "tags": []
+          }
+        }
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "BqqfluJR2-0B"
+      },
+      "source": [
+        "Actually, the projection matrix makes things worse. ***Let's NOT use it for now!!!!!***"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "metadata": {
+        "colab": {
+          "base_uri": "https://localhost:8080/"
+        },
+        "id": "pLPUbG282_SO",
+        "outputId": "57745c51-8804-4dbf-ff60-9d5b475b8440"
+      },
+      "source": [
+        "# Simple embeddings, no projection matrix added\n",
+        "all_sent_embs = encode_all_sents(all_sents, bin_model)"
+      ],
+      "execution_count": null,
+      "outputs": [
+        {
+          "output_type": "stream",
+          "text": [
+            "100%|██████████| 1220/1220 [00:22<00:00, 54.73it/s]\n"
+          ],
+          "name": "stderr"
+        }
+      ]
+    },
+    {
+      "cell_type": "code",
+      "metadata": {
+        "colab": {
+          "base_uri": "https://localhost:8080/"
+        },
+        "id": "3lvfAZQjAHvS",
+        "outputId": "41e1c935-6308-406a-bef8-014a4d941e5d"
+      },
+      "source": [
+        "all_test_embs = encode_all_sents(test_sents, bin_model)"
+      ],
+      "execution_count": null,
+      "outputs": [
+        {
+          "output_type": "stream",
+          "text": [
+            "100%|██████████| 306/306 [00:05<00:00, 54.31it/s]\n"
+          ],
+          "name": "stderr"
+        }
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "qyLt6Pl84UHc"
+      },
+      "source": [
+        "### Train classifiers"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "metadata": {
+        "id": "vCR_wG18E7aC"
+      },
+      "source": [
+        "from sklearn.model_selection import cross_val_score\n",
+        "from sklearn.metrics import classification_report"
+      ],
+      "execution_count": null,
+      "outputs": []
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "rQyQIFVEBZsk"
+      },
+      "source": [
+        "1. Let's start with Random Forests!"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "metadata": {
+        "id": "YW021S1J_JH9"
+      },
+      "source": [
+        "from sklearn.ensemble import RandomForestClassifier"
+      ],
+      "execution_count": null,
+      "outputs": []
+    },
+    {
+      "cell_type": "code",
+      "metadata": {
+        "id": "dbQ3_grn_I82"
+      },
+      "source": [
+        "clf = RandomForestClassifier(n_estimators=100, max_depth=3, random_state=69420)"
+      ],
+      "execution_count": null,
+      "outputs": []
+    },
+    {
+      "cell_type": "code",
+      "metadata": {
+        "colab": {
+          "base_uri": "https://localhost:8080/"
+        },
+        "id": "fAM7nFv3_IxT",
+        "outputId": "54c22d6a-7303-42d7-cdd3-819dad2633b8"
+      },
+      "source": [
+        "clf.fit(np.vstack(all_sent_embs), all_labels)"
+      ],
+      "execution_count": null,
+      "outputs": [
+        {
+          "output_type": "execute_result",
+          "data": {
+            "text/plain": [
+              "RandomForestClassifier(bootstrap=True, ccp_alpha=0.0, class_weight=None,\n",
+              "                       criterion='gini', max_depth=3, max_features='auto',\n",
+              "                       max_leaf_nodes=None, max_samples=None,\n",
+              "                       min_impurity_decrease=0.0, min_impurity_split=None,\n",
+              "                       min_samples_leaf=1, min_samples_split=2,\n",
+              "                       min_weight_fraction_leaf=0.0, n_estimators=100,\n",
+              "                       n_jobs=None, oob_score=False, random_state=69420,\n",
+              "                       verbose=0, warm_start=False)"
+            ]
+          },
+          "metadata": {
+            "tags": []
+          },
+          "execution_count": 136
+        }
+      ]
+    },
+    {
+      "cell_type": "code",
+      "metadata": {
+        "id": "0MsZbft8_6_1"
+      },
+      "source": [
+        "clf_preds = [clf.predict(sent_emb)[0] for sent_emb in all_test_embs]"
+      ],
+      "execution_count": null,
+      "outputs": []
+    },
+    {
+      "cell_type": "code",
+      "metadata": {
+        "colab": {
+          "base_uri": "https://localhost:8080/"
+        },
+        "id": "2uLnsvS6GVVk",
+        "outputId": "5346cb46-8f87-4a18-83d8-be9391c26027"
+      },
+      "source": [
+        " print(classification_report(test_labels, clf_preds))"
+      ],
+      "execution_count": null,
+      "outputs": [
+        {
+          "output_type": "stream",
+          "text": [
+            "               precision    recall  f1-score   support\n",
+            "\n",
+            "    Incentive       0.89      0.93      0.91       228\n",
+            "not_Incentive       0.76      0.68      0.72        78\n",
+            "\n",
+            "     accuracy                           0.86       306\n",
+            "    macro avg       0.83      0.80      0.81       306\n",
+            " weighted avg       0.86      0.86      0.86       306\n",
+            "\n"
+          ],
+          "name": "stdout"
+        }
+      ]
+    },
+    {
+      "cell_type": "code",
+      "metadata": {
+        "id": "Y_k1CWeJATgy"
+      },
+      "source": [
+        "numeric_preds = labels2numeric(clf_preds, label_names)\n",
+        "numeric_test_labels = labels2numeric(test_labels, label_names)\n",
+        "evaluator = ModelEvaluator(label_names, y_true=numeric_test_labels, y_pred=numeric_preds)"
+      ],
+      "execution_count": null,
+      "outputs": []
+    },
+    {
+      "cell_type": "code",
+      "metadata": {
+        "colab": {
+          "base_uri": "https://localhost:8080/",
+          "height": 455
+        },
+        "id": "jHI9hf6EAy_N",
+        "outputId": "1544a62f-3b4c-4aaf-b98b-82d9d3044783"
+      },
+      "source": [
+        "evaluator.plot_confusion_matrix(color_map='Blues')"
+      ],
+      "execution_count": null,
+      "outputs": [
+        {
+          "output_type": "display_data",
+          "data": {
+            "image/png": "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\n",
+            "text/plain": [
+              "<Figure size 576x432 with 2 Axes>"
+            ]
+          },
+          "metadata": {
+            "tags": []
+          }
+        }
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "yXdaChWFBrWg"
+      },
+      "source": [
+        "Honestly, without Grid Search and 5-fold Cross Validation, these are not bad results... We should add those though!"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "Y3DoYgdaDHJ6"
+      },
+      "source": [
+        "2. Now, we're gonna try Support Vector Machines"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "metadata": {
+        "id": "lcNVNbZVDBX_"
+      },
+      "source": [
+        "from sklearn import svm"
+      ],
+      "execution_count": null,
+      "outputs": []
+    },
+    {
+      "cell_type": "code",
+      "metadata": {
+        "id": "vD4-V3KpAy3k"
+      },
+      "source": [
+        "clf = svm.SVC(gamma=0.001, C=100.)"
+      ],
+      "execution_count": null,
+      "outputs": []
+    },
+    {
+      "cell_type": "code",
+      "metadata": {
+        "colab": {
+          "base_uri": "https://localhost:8080/"
+        },
+        "id": "oeqA23VGAyoK",
+        "outputId": "3d905ff1-b4e8-4d2b-8887-5a6eb5e497e1"
+      },
+      "source": [
+        "clf.fit(np.vstack(all_sent_embs), all_labels)"
+      ],
+      "execution_count": null,
+      "outputs": [
+        {
+          "output_type": "execute_result",
+          "data": {
+            "text/plain": [
+              "SVC(C=100.0, break_ties=False, cache_size=200, class_weight=None, coef0=0.0,\n",
+              "    decision_function_shape='ovr', degree=3, gamma=0.001, kernel='rbf',\n",
+              "    max_iter=-1, probability=False, random_state=None, shrinking=True,\n",
+              "    tol=0.001, verbose=False)"
+            ]
+          },
+          "metadata": {
+            "tags": []
+          },
+          "execution_count": 130
+        }
+      ]
+    },
+    {
+      "cell_type": "code",
+      "metadata": {
+        "id": "tyytG01oEC4G"
+      },
+      "source": [
+        "clf_preds = [clf.predict(sent_emb)[0] for sent_emb in all_test_embs]"
+      ],
+      "execution_count": null,
+      "outputs": []
+    },
+    {
+      "cell_type": "code",
+      "metadata": {
+        "id": "SEneBVYFD3DW"
+      },
+      "source": [
+        "numeric_preds = labels2numeric(clf_preds, label_names)\n",
+        "numeric_test_labels = labels2numeric(test_labels, label_names)\n",
+        "evaluator = ModelEvaluator(label_names, y_true=numeric_test_labels, y_pred=numeric_preds)"
+      ],
+      "execution_count": null,
+      "outputs": []
+    },
+    {
+      "cell_type": "code",
+      "metadata": {
+        "colab": {
+          "base_uri": "https://localhost:8080/"
+        },
+        "id": "cscCU733D664",
+        "outputId": "51dd0385-c7ae-44f6-f090-936b4161136a"
+      },
+      "source": [
+        "print(classification_report(test_labels, clf_preds))"
+      ],
+      "execution_count": null,
+      "outputs": [
+        {
+          "output_type": "stream",
+          "text": [
+            "               precision    recall  f1-score   support\n",
+            "\n",
+            "    Incentive       0.91      0.92      0.92       228\n",
+            "not_Incentive       0.76      0.74      0.75        78\n",
+            "\n",
+            "     accuracy                           0.88       306\n",
+            "    macro avg       0.84      0.83      0.84       306\n",
+            " weighted avg       0.87      0.88      0.88       306\n",
+            "\n"
+          ],
+          "name": "stdout"
+        }
+      ]
+    },
+    {
+      "cell_type": "code",
+      "metadata": {
+        "colab": {
+          "base_uri": "https://localhost:8080/",
+          "height": 455
+        },
+        "id": "oxYzT51aD60j",
+        "outputId": "c7af7030-d09a-43a0-c868-5ac36af22b7a"
+      },
+      "source": [
+        "evaluator.plot_confusion_matrix(color_map='Blues')"
+      ],
+      "execution_count": null,
+      "outputs": [
+        {
+          "output_type": "display_data",
+          "data": {
+            "image/png": "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\n",
+            "text/plain": [
+              "<Figure size 576x432 with 2 Axes>"
+            ]
+          },
+          "metadata": {
+            "tags": []
+          }
+        }
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "H-ki8dU5EUaq"
+      },
+      "source": [
+        "Ok, so SVMs are slightly better than Random Forests at ***differentiating*** text! There's a 1-2% decrease in performance for the incentive class, but a 6% gain in non-incentives. If these results remain when doing cross validation and grid search, then I'd recommend going for the SVMs.\n",
+        "\n",
+        "\n",
+        "**Next steps:**\n",
+        "- Add Grid Search Cross Validation from sklearn"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "metadata": {
+        "id": "9y34yQNqD6TO"
+      },
+      "source": [
+        ""
+      ],
+      "execution_count": null,
+      "outputs": []
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "mbfAvxu0-_Qi"
+      },
+      "source": [
+        "### What about... Beto?\n",
+        "I downloaded the weights and placed them in the folder below: "
+      ]
+    },
+    {
+      "cell_type": "code",
+      "metadata": {
+        "id": "n8I9IyxRrCm5",
+        "colab": {
+          "base_uri": "https://localhost:8080/"
+        },
+        "outputId": "82ed8f16-16d3-4b71-c0e2-31304ae96327"
+      },
+      "source": [
+        "!pip install transformers"
+      ],
+      "execution_count": null,
+      "outputs": [
+        {
+          "output_type": "stream",
+          "text": [
+            "Requirement already satisfied: transformers in /usr/local/lib/python3.6/dist-packages (4.2.1)\n",
+            "Requirement already satisfied: importlib-metadata; python_version < \"3.8\" in /usr/local/lib/python3.6/dist-packages (from transformers) (3.3.0)\n",
+            "Requirement already satisfied: regex!=2019.12.17 in /usr/local/lib/python3.6/dist-packages (from transformers) (2019.12.20)\n",
+            "Requirement already satisfied: dataclasses; python_version < \"3.7\" in /usr/local/lib/python3.6/dist-packages (from transformers) (0.8)\n",
+            "Requirement already satisfied: numpy in /usr/local/lib/python3.6/dist-packages (from transformers) (1.19.5)\n",
+            "Requirement already satisfied: tqdm>=4.27 in /usr/local/lib/python3.6/dist-packages (from transformers) (4.41.1)\n",
+            "Requirement already satisfied: tokenizers==0.9.4 in /usr/local/lib/python3.6/dist-packages (from transformers) (0.9.4)\n",
+            "Requirement already satisfied: filelock in /usr/local/lib/python3.6/dist-packages (from transformers) (3.0.12)\n",
+            "Requirement already satisfied: sacremoses in /usr/local/lib/python3.6/dist-packages (from transformers) (0.0.43)\n",
+            "Requirement already satisfied: packaging in /usr/local/lib/python3.6/dist-packages (from transformers) (20.8)\n",
+            "Requirement already satisfied: requests in /usr/local/lib/python3.6/dist-packages (from transformers) (2.23.0)\n",
+            "Requirement already satisfied: typing-extensions>=3.6.4; python_version < \"3.8\" in /usr/local/lib/python3.6/dist-packages (from importlib-metadata; python_version < \"3.8\"->transformers) (3.7.4.3)\n",
+            "Requirement already satisfied: zipp>=0.5 in /usr/local/lib/python3.6/dist-packages (from importlib-metadata; python_version < \"3.8\"->transformers) (3.4.0)\n",
+            "Requirement already satisfied: joblib in /usr/local/lib/python3.6/dist-packages (from sacremoses->transformers) (1.0.0)\n",
+            "Requirement already satisfied: click in /usr/local/lib/python3.6/dist-packages (from sacremoses->transformers) (7.1.2)\n",
+            "Requirement already satisfied: six in /usr/local/lib/python3.6/dist-packages (from sacremoses->transformers) (1.15.0)\n",
+            "Requirement already satisfied: pyparsing>=2.0.2 in /usr/local/lib/python3.6/dist-packages (from packaging->transformers) (2.4.7)\n",
+            "Requirement already satisfied: urllib3!=1.25.0,!=1.25.1,<1.26,>=1.21.1 in /usr/local/lib/python3.6/dist-packages (from requests->transformers) (1.24.3)\n",
+            "Requirement already satisfied: idna<3,>=2.5 in /usr/local/lib/python3.6/dist-packages (from requests->transformers) (2.10)\n",
+            "Requirement already satisfied: certifi>=2017.4.17 in /usr/local/lib/python3.6/dist-packages (from requests->transformers) (2020.12.5)\n",
+            "Requirement already satisfied: chardet<4,>=3.0.2 in /usr/local/lib/python3.6/dist-packages (from requests->transformers) (3.0.4)\n"
+          ],
+          "name": "stdout"
+        }
+      ]
+    },
+    {
+      "cell_type": "code",
+      "metadata": {
+        "id": "mxk0ZHUFHnkk"
+      },
+      "source": [
+        "beto_path = f\"{base_path}/Modeling/BETO/pytorch/\""
+      ],
+      "execution_count": null,
+      "outputs": []
+    },
+    {
+      "cell_type": "code",
+      "metadata": {
+        "id": "7lHUcXtY3PKP"
+      },
+      "source": [
+        "from transformers import BertTokenizer, BertForSequenceClassification"
+      ],
+      "execution_count": null,
+      "outputs": []
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "TI9IF0UmKYZE"
+      },
+      "source": [
+        "***IMPORTANT:*** I was not able to figure out a way of using the fine tuning results from the models above so I'm gonna use BETO out of the box, for both encoding/classification and see how it goes.\n",
+        "\n",
+        "The following cells are a demo of how the model should be put to use - once you understand it, feel free to skip this part!"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "metadata": {
+        "colab": {
+          "base_uri": "https://localhost:8080/"
+        },
+        "id": "kmJ_mpC37ICQ",
+        "outputId": "f350a9e6-e3a8-476e-846f-43b387302ce1"
+      },
+      "source": [
+        "tokenizer = BertTokenizer.from_pretrained(beto_path)\n",
+        "model = BertForSequenceClassification.from_pretrained(beto_path)"
+      ],
+      "execution_count": null,
+      "outputs": [
+        {
+          "output_type": "stream",
+          "text": [
+            "Some weights of the model checkpoint at /content/drive/MyDrive/WRI-LatinAmerica-Talent/Modeling/BETO/pytorch/ were not used when initializing BertForSequenceClassification: ['cls.predictions.bias', 'cls.predictions.transform.dense.weight', 'cls.predictions.transform.dense.bias', 'cls.predictions.transform.LayerNorm.weight', 'cls.predictions.transform.LayerNorm.bias', 'cls.predictions.decoder.weight', 'cls.seq_relationship.weight', 'cls.seq_relationship.bias']\n",
+            "- This IS expected if you are initializing BertForSequenceClassification from the checkpoint of a model trained on another task or with another architecture (e.g. initializing a BertForSequenceClassification model from a BertForPreTraining model).\n",
+            "- This IS NOT expected if you are initializing BertForSequenceClassification from the checkpoint of a model that you expect to be exactly identical (initializing a BertForSequenceClassification model from a BertForSequenceClassification model).\n",
+            "Some weights of BertForSequenceClassification were not initialized from the model checkpoint at /content/drive/MyDrive/WRI-LatinAmerica-Talent/Modeling/BETO/pytorch/ and are newly initialized: ['classifier.weight', 'classifier.bias']\n",
+            "You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.\n"
+          ],
+          "name": "stderr"
+        }
+      ]
+    },
+    {
+      "cell_type": "code",
+      "metadata": {
+        "id": "EGRL66pBIzuy"
+      },
+      "source": [
+        "classes = [\"no es parafrasis\", \"es parafrasis\"]\n",
+        "\n",
+        "sequence_0 = \"La compañia Hugging esta basada en Nueva York\"\n",
+        "sequence_1 = \"Las manzanas son malas para la salud\"\n",
+        "sequence_2 = \"La sede principal de Hugging esta en Manhattan\"\n",
+        "\n",
+        "paraphrase = tokenizer(sequence_0, sequence_2, return_tensors=\"pt\")\n",
+        "not_paraphrase = tokenizer(sequence_0, sequence_1, return_tensors=\"pt\")\n",
+        "\n",
+        "paraphrase_classification_logits = model(**paraphrase).logits\n",
+        "not_paraphrase_classification_logits = model(**not_paraphrase).logits\n",
+        "\n",
+        "paraphrase_results = torch.softmax(paraphrase_classification_logits, dim=1).tolist()[0]\n",
+        "not_paraphrase_results = torch.softmax(not_paraphrase_classification_logits, dim=1).tolist()[0]"
+      ],
+      "execution_count": null,
+      "outputs": []
+    },
+    {
+      "cell_type": "code",
+      "metadata": {
+        "colab": {
+          "base_uri": "https://localhost:8080/"
+        },
+        "id": "C-Pkqe-vI3ld",
+        "outputId": "5d208439-c5ae-4da4-b6a1-e748c3a2f839"
+      },
+      "source": [
+        "print(\">>> Deberia de ser parafrasis:\")\n",
+        "for i in range(len(classes)):\n",
+        "  print(f\"{classes[i]}: {int(round(paraphrase_results[i] * 100))}%\")\n",
+        "\n",
+        "print(\"\\n>>> NO deberia de ser parafrasis:\")\n",
+        "for i in range(len(classes)):\n",
+        "  print(f\"{classes[i]}: {int(round(not_paraphrase_results[i] * 100))}%\")"
+      ],
+      "execution_count": null,
+      "outputs": [
+        {
+          "output_type": "stream",
+          "text": [
+            ">>> Deberia de ser parafrasis:\n",
+            "no es parafrasis: 41%\n",
+            "es parafrasis: 59%\n",
+            "\n",
+            ">>> NO deberia de ser parafrasis:\n",
+            "no es parafrasis: 61%\n",
+            "es parafrasis: 39%\n"
+          ],
+          "name": "stdout"
+        }
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "0dRr9sJQKezT"
+      },
+      "source": [
+        "OK! Now it's time to apply it to our data. We will try it out with our test set, just to have a fair comparison"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "metadata": {
+        "id": "ybHMwJE-JyvG"
+      },
+      "source": [
+        "tokenized_sents = tokenizer(test_sents, padding=True, return_tensors=\"pt\")\n",
+        "clf_logits = model(**tokenized_sents).logits\n",
+        "clf_results = torch.softmax(clf_logits, dim=1).tolist()[0]"
+      ],
+      "execution_count": null,
+      "outputs": []
+    },
+    {
+      "cell_type": "code",
+      "metadata": {
+        "id": "wELlB7iUJ1T3"
+      },
+      "source": [
+        "# This stores the index of the highest score - in other words, our label\n",
+        "clf_preds = [np.argmax(logits) for logits in clf_results]"
+      ],
+      "execution_count": null,
+      "outputs": []
+    },
+    {
+      "cell_type": "code",
+      "metadata": {
+        "id": "GnrG0AwcM-G5"
+      },
+      "source": [
+        "print(classification_report(test_labels, clf_preds))"
+      ],
+      "execution_count": null,
+      "outputs": []
+    },
+    {
+      "cell_type": "code",
+      "metadata": {
+        "colab": {
+          "base_uri": "https://localhost:8080/"
+        },
+        "id": "qdV91B4IMZdz",
+        "outputId": "c6903dcb-c967-428f-a157-42717317da82"
+      },
+      "source": [
+        "evaluator = ModelEvaluator(label_names, y_true=numeric_test_labels, y_pred=clf_preds)\n",
+        "evaluator.plot_confusion_matrix(color_map='Blues')"
+      ],
+      "execution_count": null,
+      "outputs": [
+        {
+          "output_type": "execute_result",
+          "data": {
+            "text/plain": [
+              "1"
+            ]
+          },
+          "metadata": {
+            "tags": []
+          },
+          "execution_count": 161
+        }
+      ]
+    },
+    {
+      "cell_type": "code",
+      "metadata": {
+        "id": "eMEJ67HrMYtN"
+      },
+      "source": [
+        ""
+      ],
+      "execution_count": null,
+      "outputs": []
     }
-   ],
-   "source": [
-    "evaluator = ModelEvaluator(label_names, y_true=numeric_test_labels, y_pred=clf_preds)\n",
-    "evaluator.plot_confusion_matrix(color_map='Blues')"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {
-    "id": "0cuHUa-MMZJR"
-   },
-   "outputs": [],
-   "source": []
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {
-    "id": "eMEJ67HrMYtN"
-   },
-   "outputs": [],
-   "source": []
-  }
- ],
- "metadata": {
-  "accelerator": "GPU",
-  "colab": {
-   "collapsed_sections": [],
-   "name": "BinaryClassifierGoogleColab.ipynb",
-   "provenance": []
-  },
-  "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": "3.7.9"
-  },
-  "toc": {
-   "base_numbering": 1,
-   "nav_menu": {},
-   "number_sections": true,
-   "sideBar": true,
-   "skip_h1_title": false,
-   "title_cell": "Table of Contents",
-   "title_sidebar": "Contents",
-   "toc_cell": false,
-   "toc_position": {},
-   "toc_section_display": true,
-   "toc_window_display": false
-  }
- },
- "nbformat": 4,
- "nbformat_minor": 4
-}
+  ]
+}
\ No newline at end of file