diff --git a/docs/guides/weak-supervision.ipynb b/docs/guides/weak-supervision.ipynb
index f227c68f4db..70411174449 100644
--- a/docs/guides/weak-supervision.ipynb
+++ b/docs/guides/weak-supervision.ipynb
@@ -33,9 +33,10 @@
     "\n",
     "### Weak labeling using the UI\n",
     "\n",
-    "Since version 0.8.0 you can find and define rules directly in the UI. The [Define rules mode](../reference/webapp/define_rules.md) is found below the [Annotate mode](../reference/webapp/annotate_records.md) on the right sidebar.\n",
-    "\n",
-    "The video below shows how you can interactively find and save rules with the UI. For more a full example check the Weak supervision tutorial.\n",
+    "Since version 0.8.0 you can find and define rules directly in the UI. \n",
+    "The [Define rules mode](../reference/webapp/define_rules.md) is found in the right side bar of the [Dataset page](../reference/webapp/dataset.md).\n",
+    "The video below shows how you can interactively find and save rules with the UI. \n",
+    "For a full example check the [Weak supervision tutorial](../tutorials/weak-supervision-with-rubrix.ipynb).\n",
     "\n",
     "<video width=\"100%\" controls><source src=\"../_static/tutorials/weak-supervision-with-rubrix/ws_news.mp4\" type=\"video/mp4\"></video>\n",
     "\n",
@@ -112,7 +113,7 @@
     "4. Once you are satisfied with your weak labels, use the matrix of the `WeakLabels` instance with your library/method of choice to build a training set or even train a downstream text classification model.\n",
     "\n",
     "\n",
-    "This guide shows you an end-to-end example using Snorkel and Flyingsquid. Let's get started!"
+    "This guide shows you an end-to-end example using Snorkel, Flyingsquid and Weasel. Let's get started!"
    ]
   },
   {
@@ -127,7 +128,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 11,
+   "execution_count": 2,
    "id": "6b1e00af-c6f9-42aa-81aa-2976c9591b36",
    "metadata": {},
    "outputs": [
@@ -211,15 +212,22 @@
        "</div>"
       ],
       "text/plain": [
-       "   Unnamed: 0            author                 date                                               text  label  video\n",
-       "0           0  Alessandro leite  2014-11-05T22:21:36  pls http://www10.vakinha.com.br/VaquinhaE.aspx...   -1.0      1\n",
-       "1           1      Salim Tayara  2014-11-02T14:33:30  if your like drones, plz subscribe to Kamal Ta...   -1.0      1\n",
-       "2           2           Phuc Ly  2014-01-20T15:27:47                     go here to check the views :3   -1.0      1\n",
-       "3           3      DropShotSk8r  2014-01-19T04:27:18            Came here to check the views, goodbye.   -1.0      1\n",
-       "4           4            css403  2014-11-07T14:25:48                      i am 2,126,492,636 viewer :D   -1.0      1"
+       "   Unnamed: 0            author                 date  \\\n",
+       "0           0  Alessandro leite  2014-11-05T22:21:36   \n",
+       "1           1      Salim Tayara  2014-11-02T14:33:30   \n",
+       "2           2           Phuc Ly  2014-01-20T15:27:47   \n",
+       "3           3      DropShotSk8r  2014-01-19T04:27:18   \n",
+       "4           4            css403  2014-11-07T14:25:48   \n",
+       "\n",
+       "                                                text  label  video  \n",
+       "0  pls http://www10.vakinha.com.br/VaquinhaE.aspx...   -1.0      1  \n",
+       "1  if your like drones, plz subscribe to Kamal Ta...   -1.0      1  \n",
+       "2                     go here to check the views :3   -1.0      1  \n",
+       "3            Came here to check the views, goodbye.   -1.0      1  \n",
+       "4                      i am 2,126,492,636 viewer :D   -1.0      1  "
       ]
      },
-     "execution_count": 11,
+     "execution_count": 2,
      "metadata": {},
      "output_type": "execute_result"
     }
@@ -283,7 +291,7 @@
    "id": "baafda4f-45c0-49d6-9c37-7473c6888ebe",
    "metadata": {},
    "source": [
-    "After this step, you have a fully browsable dataset available at `http://localhost:6900/weak_supervision_yt` (or the base URL where your Rubrix instance is hosted)."
+    "After this step, you have a fully browsable dataset available that you can access via the [Rubrix web app](../reference/webapp/index.md)."
    ]
   },
   {
@@ -293,7 +301,9 @@
    "source": [
     "## 2. Defining rules\n",
     "\n",
-    "Let's now define some of the rules proposed in the tutorial [Snorkel Intro Tutorial: Data Labeling](https://www.snorkel.org/use-cases/01-spam-tutorial). Most of these rules can be defined directly in the UI using [Elasticsearch's query string DSL](../reference/webapp/search_records.md). \n",
+    "Let's now define some of the rules proposed in the tutorial [Snorkel Intro Tutorial: Data Labeling](https://www.snorkel.org/use-cases/01-spam-tutorial). \n",
+    "Most of these rules can be defined directly with our web app in the [Define rules mode](../reference/webapp/define_rules.md) and [Elasticsearch's query strings](../reference/webapp/search_records.md). \n",
+    "Afterward, you can conveniently load them into your notebook with the [load_rules function](../reference/python/python_labeling.rst#rubrix.labeling.text_classification.rule.load_rules).\n",
     "\n",
     "Rules can also be defined programmatically as shown below. Depending on your use case and team structure you can mix and match both interfaces (UI or Python).\n",
     "\n",
@@ -571,6 +581,231 @@
     "Let's see some examples:"
    ]
   },
+  {
+   "cell_type": "markdown",
+   "id": "1fb78718",
+   "metadata": {},
+   "source": [
+    "### A simple majority vote\n",
+    "\n",
+    "As a first example we will show you, how to use the `WeakLabels` object together with a simple majority vote model.\n",
+    "For this we will take the implementation by Snorkel, and evaluate it with the help of sklearn's metrics module."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "f3ba07e7",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "%pip install snorkel scikit-learn -qqq"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "08eb2b66",
+   "metadata": {},
+   "source": [
+    "The majority vote model is arguably the most straightforward label model.\n",
+    "On a per-record basis, it simply counts the votes for each label returned by the rules, and takes the majority vote.\n",
+    "Snorkel provides a neat implementation of this logic in its `MajorityLabelVoter`."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "15888261",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from snorkel.labeling.model import MajorityLabelVoter\n",
+    "\n",
+    "# instantiate the majority vote label model\n",
+    "majority_model = MajorityLabelVoter()"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "293d163f-ba08-424c-93de-ef0420531ca9",
+   "metadata": {},
+   "source": [
+    "Let's go on and evaluate this baseline.\n",
+    "To break ties when there is no majority vote, we choose the _\"random\"_ policy that randomly selects one of the tied labels. \n",
+    "In this way we avoid a bias towards label models that produce fewer but more certain weak labels, and makes the comparison between the different label models fairer."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "1d588f00-c8e1-4bec-8e99-bd78639257b5",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# compute accuracy\n",
+    "majority_model.score(\n",
+    "    L=weak_labels.matrix(has_annotation=True),\n",
+    "    Y=weak_labels.annotation(),\n",
+    "    tie_break_policy=\"random\",\n",
+    ")\n",
+    "# {'accuracy': 0.844}"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "a619247f-3d9d-44ae-bb9f-d07c9120c1dc",
+   "metadata": {},
+   "source": [
+    "As we will see further down, an accuracy of 0.844 is a very decent baseline.\n",
+    "Choosing to simply ignore tiebreaks and abstentions (by setting the tiebreak policy to _\"abstain\"_), we would obtain an accuracy of nearly 0.96.\n",
+    "\n",
+    "When predicting weak labels to train a down-stream model, you probably want to discard the abstentions and tiebreaks."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "b8de05c3-84ef-40de-8f73-0157a4aa1074",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# get predictions for training a down-stream model\n",
+    "predictions = majority_model.predict(L=weak_labels.matrix(has_annotation=False))\n",
+    "\n",
+    "# records for training\n",
+    "training_records = weak_labels.records(has_annotation=False)\n",
+    "\n",
+    "# mask to ignore abstentions/tiebreaks\n",
+    "idx = predictions != -1\n",
+    "\n",
+    "# combine records and predictions\n",
+    "training_data = pd.DataFrame(\n",
+    "    [\n",
+    "        {\"text\": rec.inputs[\"text\"], \"label\": weak_labels.int2label[label]} \n",
+    "        for rec, label in zip(training_records, predictions)\n",
+    "    ]\n",
+    ")[idx]"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 240,
+   "id": "39e06fd0-caa6-4707-a667-030b52ad4be9",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>text</th>\n",
+       "      <th>label</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>0</th>\n",
+       "      <td>Hey I&amp;#39;m a British youtuber!!&lt;br /&gt;I upload...</td>\n",
+       "      <td>SPAM</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1</th>\n",
+       "      <td>NOKIA spotted</td>\n",
+       "      <td>HAM</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2</th>\n",
+       "      <td>Dance :)</td>\n",
+       "      <td>HAM</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>3</th>\n",
+       "      <td>You guys should check out this EXTRAORDINARY w...</td>\n",
+       "      <td>SPAM</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>4</th>\n",
+       "      <td>Need money ? check my channel and subscribe,so...</td>\n",
+       "      <td>SPAM</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>...</th>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1579</th>\n",
+       "      <td>Please check out my acoustic cover channel :) ...</td>\n",
+       "      <td>SPAM</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1580</th>\n",
+       "      <td>PLEASE SUBSCRIBE ME!!!!!!!!!!!!!!!!!!!!!!!!!!!...</td>\n",
+       "      <td>SPAM</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1581</th>\n",
+       "      <td>&lt;a href=\"http://www.gofundme.com/Helpmypitbull...</td>\n",
+       "      <td>SPAM</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1582</th>\n",
+       "      <td>I love this song so much!:-D I've heard it so ...</td>\n",
+       "      <td>HAM</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1585</th>\n",
+       "      <td>Check out this video on YouTube:</td>\n",
+       "      <td>SPAM</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "<p>1055 rows × 2 columns</p>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "                                                   text label\n",
+       "0     Hey I&#39;m a British youtuber!!<br />I upload...  SPAM\n",
+       "1                                        NOKIA spotted   HAM\n",
+       "2                                             Dance :)   HAM\n",
+       "3     You guys should check out this EXTRAORDINARY w...  SPAM\n",
+       "4     Need money ? check my channel and subscribe,so...  SPAM\n",
+       "...                                                 ...   ...\n",
+       "1579  Please check out my acoustic cover channel :) ...  SPAM\n",
+       "1580  PLEASE SUBSCRIBE ME!!!!!!!!!!!!!!!!!!!!!!!!!!!...  SPAM\n",
+       "1581  <a href=\"http://www.gofundme.com/Helpmypitbull...  SPAM\n",
+       "1582  I love this song so much!:-D I've heard it so ...   HAM\n",
+       "1585                  Check out this video on YouTube:  SPAM\n",
+       "\n",
+       "[1055 rows x 2 columns]"
+      ]
+     },
+     "execution_count": 240,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "# preview training data\n",
+    "training_data"
+   ]
+  },
   {
    "cell_type": "markdown",
    "id": "2186a612-57a4-4118-97fb-927ed87df96d",
@@ -578,7 +813,7 @@
    "source": [
     "### Label model with Snorkel\n",
     "\n",
-    "Snorkel is by far the most popular option for using weak supervision, and Rubrix provides built-in support for it. \n",
+    "Snorkel's label model is by far the most popular option for using weak supervision, and Rubrix provides built-in support for it. \n",
     "Using Snorkel with Rubrix's `WeakLabels` is as simple as:"
    ]
   },
@@ -602,13 +837,30 @@
     "from rubrix.labeling.text_classification import Snorkel\n",
     "\n",
     "# we pass our WeakLabels instance to our Snorkel label model\n",
-    "label_model = Snorkel(weak_labels)\n",
+    "snorkel_model = Snorkel(weak_labels)\n",
     "\n",
     "# we fit the model\n",
-    "label_model.fit()\n",
-    "\n",
+    "snorkel_model.fit(lr=0.001, n_epochs=50)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "b5f6346d-1a34-4efa-87de-f3e69b7550ca",
+   "metadata": {},
+   "source": [
+    "When fitting the snorkel model, we recommend performing a quick grid search for the learning rate `lr` and the number of epochs `n_epochs`."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "9c82339f-561d-4c59-89ee-c49ddbe8da83",
+   "metadata": {},
+   "outputs": [],
+   "source": [
     "# we check its performance\n",
-    "label_model.score()"
+    "snorkel_model.score(tie_break_policy=\"abstain\")\n",
+    "# {'accuracy': 0.848, ...}"
    ]
   },
   {
@@ -616,8 +868,9 @@
    "id": "075463f3-8695-4d28-af24-8860b2efd691",
    "metadata": {},
    "source": [
-    "After fitting your label model, you can quickly explore its predictions, before building a training set for training a downstream text classifier. \n",
+    "When choosing to simply ignore tiebreaks and abstentions in the score (by setting the tiebreak policy to _\"abstain\"_), we would obtain an accuracy of about 0.95.\n",
     "\n",
+    "After fitting your label model, you can quickly explore its predictions, before building a training set for training a downstream text classifier. \n",
     "This step is useful for validation, manual revision, or defining score thresholds for accepting labels from your label model (for example, only considering labels with a score greater then 0.8.)"
    ]
   },
@@ -629,10 +882,151 @@
    "outputs": [],
    "source": [
     "# get your training records with the predictions of the label model\n",
-    "records_for_training = label_model.predict()\n",
+    "records_for_training = snorkel_model.predict()\n",
     "\n",
-    "# log the records to a new dataset in Rubrix\n",
-    "rb.log(records_for_training, name=\"snorkel_results\")"
+    "# optional: log the records to a new dataset in Rubrix\n",
+    "rb.log(records_for_training, name=\"snorkel_results\")\n",
+    "\n",
+    "# extract training data\n",
+    "training_data = pd.DataFrame(\n",
+    "    [\n",
+    "        {\"text\": rec.inputs[\"text\"], \"label\": rec.prediction[0][0]} \n",
+    "        for rec in records_for_training\n",
+    "    ]\n",
+    ")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 225,
+   "id": "4e18b365-5412-47c7-b0e6-e140bd0c2dc6",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>text</th>\n",
+       "      <th>label</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>0</th>\n",
+       "      <td>Hey I&amp;#39;m a British youtuber!!&lt;br /&gt;I upload...</td>\n",
+       "      <td>SPAM</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1</th>\n",
+       "      <td>NOKIA spotted</td>\n",
+       "      <td>HAM</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2</th>\n",
+       "      <td>Dance :)</td>\n",
+       "      <td>HAM</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>3</th>\n",
+       "      <td>You guys should check out this EXTRAORDINARY w...</td>\n",
+       "      <td>SPAM</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>4</th>\n",
+       "      <td>Need money ? check my channel and subscribe,so...</td>\n",
+       "      <td>SPAM</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>...</th>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1172</th>\n",
+       "      <td>Please check out my acoustic cover channel :) ...</td>\n",
+       "      <td>SPAM</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1173</th>\n",
+       "      <td>PLEASE SUBSCRIBE ME!!!!!!!!!!!!!!!!!!!!!!!!!!!...</td>\n",
+       "      <td>SPAM</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1174</th>\n",
+       "      <td>&lt;a href=\"http://www.gofundme.com/Helpmypitbull...</td>\n",
+       "      <td>SPAM</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1175</th>\n",
+       "      <td>I love this song so much!:-D I've heard it so ...</td>\n",
+       "      <td>HAM</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1176</th>\n",
+       "      <td>Check out this video on YouTube:</td>\n",
+       "      <td>SPAM</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "<p>1177 rows × 2 columns</p>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "                                                   text label\n",
+       "0     Hey I&#39;m a British youtuber!!<br />I upload...  SPAM\n",
+       "1                                        NOKIA spotted   HAM\n",
+       "2                                             Dance :)   HAM\n",
+       "3     You guys should check out this EXTRAORDINARY w...  SPAM\n",
+       "4     Need money ? check my channel and subscribe,so...  SPAM\n",
+       "...                                                 ...   ...\n",
+       "1172  Please check out my acoustic cover channel :) ...  SPAM\n",
+       "1173  PLEASE SUBSCRIBE ME!!!!!!!!!!!!!!!!!!!!!!!!!!!...  SPAM\n",
+       "1174  <a href=\"http://www.gofundme.com/Helpmypitbull...  SPAM\n",
+       "1175  I love this song so much!:-D I've heard it so ...   HAM\n",
+       "1176                  Check out this video on YouTube:  SPAM\n",
+       "\n",
+       "[1177 rows x 2 columns]"
+      ]
+     },
+     "execution_count": 225,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "# preview training data\n",
+    "training_data"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "e3f0123c-a7a5-4a35-82fc-a5a616cd5000",
+   "metadata": {},
+   "source": [
+    "<div class=\"alert alert-info\">\n",
+    "\n",
+    "Note\n",
+    "\n",
+    "For an example of how to use the `WeakLabels` object with Snorkel's raw `LabelModel` class, you can check out the [WeakLabels reference](../reference/python/python_labeling.rst#rubrix.labeling.text_classification.weak_labels.WeakLabels).\n",
+    "    \n",
+    "</div>"
    ]
   },
   {
@@ -667,13 +1061,22 @@
     "from rubrix.labeling.text_classification import FlyingSquid\n",
     "\n",
     "# we pass our WeakLabels instance to our FlyingSquid label model\n",
-    "label_model = FlyingSquid(weak_labels)\n",
+    "flyingsquid_model = FlyingSquid(weak_labels)\n",
     "\n",
     "# we fit the model\n",
-    "label_model.fit()\n",
-    "\n",
+    "flyingsquid_model.fit()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "2d81dbf5-edf9-4c8c-a876-5fcb16d73090",
+   "metadata": {},
+   "outputs": [],
+   "source": [
     "# we check its performance\n",
-    "label_model.score()"
+    "flyingsquid_model.score(tie_break_policy=\"random\")\n",
+    "# {'accuracy': 0.832, ...}"
    ]
   },
   {
@@ -681,8 +1084,9 @@
    "id": "d282da18-cc57-437a-bfd1-c13a9ac1aec4",
    "metadata": {},
    "source": [
-    "After fitting your label model, you can quickly explore its predictions, before building a training set for training a downstream text classifier. \n",
+    "When choosing to simply ignore tiebreaks and abstentions in the score (by setting the tiebreak policy to _\"abstain\"_), we would obtain an accuracy of about 0.93.\n",
     "\n",
+    "After fitting your label model, you can quickly explore its predictions, before building a training set for training a downstream text classifier. \n",
     "This step is useful for validation, manual revision, or defining score thresholds for accepting labels from your label model (for example, only considering labels with a score greater then 0.8.)"
    ]
   },
@@ -694,10 +1098,137 @@
    "outputs": [],
    "source": [
     "# get your training records with the predictions of the label model\n",
-    "records_for_training = label_model.predict()\n",
+    "records_for_training = flyingsquid_model.predict()\n",
     "\n",
     "# log the records to a new dataset in Rubrix\n",
-    "rb.log(records_for_training, name=\"flyingsquid_results\")"
+    "rb.log(records_for_training, name=\"flyingsquid_results\")\n",
+    "\n",
+    "# extract training data\n",
+    "training_data = pd.DataFrame(\n",
+    "    [\n",
+    "        {\"text\": rec.inputs[\"text\"], \"label\": rec.prediction[0][0]} \n",
+    "        for rec in records_for_training\n",
+    "    ]\n",
+    ")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 231,
+   "id": "0d641340-f82b-4af8-b86b-7c06eaf59f61",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>text</th>\n",
+       "      <th>label</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>0</th>\n",
+       "      <td>Hey I&amp;#39;m a British youtuber!!&lt;br /&gt;I upload...</td>\n",
+       "      <td>SPAM</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1</th>\n",
+       "      <td>NOKIA spotted</td>\n",
+       "      <td>HAM</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2</th>\n",
+       "      <td>Dance :)</td>\n",
+       "      <td>HAM</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>3</th>\n",
+       "      <td>You guys should check out this EXTRAORDINARY w...</td>\n",
+       "      <td>SPAM</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>4</th>\n",
+       "      <td>Need money ? check my channel and subscribe,so...</td>\n",
+       "      <td>SPAM</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>...</th>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1172</th>\n",
+       "      <td>Please check out my acoustic cover channel :) ...</td>\n",
+       "      <td>SPAM</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1173</th>\n",
+       "      <td>PLEASE SUBSCRIBE ME!!!!!!!!!!!!!!!!!!!!!!!!!!!...</td>\n",
+       "      <td>SPAM</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1174</th>\n",
+       "      <td>&lt;a href=\"http://www.gofundme.com/Helpmypitbull...</td>\n",
+       "      <td>SPAM</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1175</th>\n",
+       "      <td>I love this song so much!:-D I've heard it so ...</td>\n",
+       "      <td>HAM</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1176</th>\n",
+       "      <td>Check out this video on YouTube:</td>\n",
+       "      <td>SPAM</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "<p>1177 rows × 2 columns</p>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "                                                   text label\n",
+       "0     Hey I&#39;m a British youtuber!!<br />I upload...  SPAM\n",
+       "1                                        NOKIA spotted   HAM\n",
+       "2                                             Dance :)   HAM\n",
+       "3     You guys should check out this EXTRAORDINARY w...  SPAM\n",
+       "4     Need money ? check my channel and subscribe,so...  SPAM\n",
+       "...                                                 ...   ...\n",
+       "1172  Please check out my acoustic cover channel :) ...  SPAM\n",
+       "1173  PLEASE SUBSCRIBE ME!!!!!!!!!!!!!!!!!!!!!!!!!!!...  SPAM\n",
+       "1174  <a href=\"http://www.gofundme.com/Helpmypitbull...  SPAM\n",
+       "1175  I love this song so much!:-D I've heard it so ...   HAM\n",
+       "1176                  Check out this video on YouTube:  SPAM\n",
+       "\n",
+       "[1177 rows x 2 columns]"
+      ]
+     },
+     "execution_count": 231,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "# preview training data\n",
+    "training_data"
    ]
   },
   {
@@ -897,7 +1428,8 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "trainer.test()  # List of test metrics"
+    "trainer.test()\n",
+    "# {'accuracy': 0.94, ...}"
    ]
   },
   {