diff --git a/docs/deep_learning/evaluation_metrics.md b/docs/deep_learning/evaluation_metrics.md
index 5748b50..074aa20 100644
--- a/docs/deep_learning/evaluation_metrics.md
+++ b/docs/deep_learning/evaluation_metrics.md
@@ -115,6 +115,14 @@ $$
 \text{recall} = \frac{\text{TP}}{\text{TP} + \text{FN}}
 $$
 
+```{note}
+Precision is about your prediction.
+Recall is about reality.
+
+If your job is to identify thieves.
+```
+
+
 #### False Negative Rate
 False Negative Rate (FNR) tells us what proportion of the positive class got incorrectly classified by the classifier.
 
diff --git a/docs/index.rst b/docs/index.rst
index 1b4a156..66dc97a 100644
--- a/docs/index.rst
+++ b/docs/index.rst
@@ -93,6 +93,7 @@ Contents
    torch/pytorch_fundamental.ipynb
    torch/pytorch_workflow.ipynb
    torch/pytorch_neural_network_classification.ipynb
+   torch/pytorch_transformers.ipynb
 
 .. toctree::
    :caption: Recommendation System
diff --git a/docs/requirements.txt b/docs/requirements.txt
index 05ad6ce..4cb3ebf 100644
--- a/docs/requirements.txt
+++ b/docs/requirements.txt
@@ -1,12 +1,12 @@
-torch~=2.0
-seaborn~=0.12
-scipy~=1.10
-scikit-learn~=1.2
-torchmetrics~=0.11
-myst-nb~=0.17
-sphinx-design~=0.3
+torch~=2.2
+seaborn~=0.13
+scipy~=1.12
+scikit-learn~=1.4
+torchmetrics~=1.3
+myst-nb~=1.0
+sphinx-design~=0.5
 sphinx-copybutton
 xlrd~=2.0
 scikit-surprise~=1.1
-networkx~=3.1
+networkx~=3.2
 gensim~=4.3
\ No newline at end of file
diff --git a/docs/torch/pytorch_fundamental.ipynb b/docs/torch/pytorch_fundamental.ipynb
index e152f38..8b127fe 100644
--- a/docs/torch/pytorch_fundamental.ipynb
+++ b/docs/torch/pytorch_fundamental.ipynb
@@ -12,22 +12,23 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 1,
+   "execution_count": 2,
    "metadata": {},
    "outputs": [
     {
      "data": {
       "text/plain": [
-       "'2.0.0'"
+       "'2.1.0'"
       ]
      },
-     "execution_count": 1,
+     "execution_count": 2,
      "metadata": {},
      "output_type": "execute_result"
     }
    ],
    "source": [
     "import torch\n",
+    "import numpy as np\n",
     "\n",
     "torch.__version__"
    ]
@@ -1734,6 +1735,90 @@
     "x[0, 0, :] # same as x[0][0]"
    ]
   },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## Pytorch Best Practise"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "You can also use torch.as_tensor() to convert a numpy array to a torch tensor, This will not create a new copy of the data"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 5,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "a = np.random.rand(3, 3)\n",
+    "# Bad way\n",
+    "t1 = torch. tensor(a)\n",
+    "\n",
+    "# Good way\n",
+    "t2 = torch.as_tensor(a)\n",
+    "t3 = torch.from_numpy(a)\n"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "Avoid cpu, item() these will use functions to tranfer data between devices"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 6,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "tensor([[0.7461, 0.3134],\n",
+       "        [0.4447, 0.7793]])"
+      ]
+     },
+     "execution_count": 6,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "t= torch.rand(2,2)\n",
+    "# bad way\n",
+    "t.cpu ()\n",
+    "t[0][0].item()\n",
+    "t. numpy ()\n",
+    "\n",
+    "# good way\n",
+    "t.detach ()"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "Create tensor direclty on GPU"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 9,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# bad way\n",
+    "# t = torch.rand(2,2).cuda()\n",
+    "\n",
+    "# good way\n",
+    "# t = torch. rand(2,2, device=\"cuda\")"
+   ]
+  },
   {
    "cell_type": "code",
    "execution_count": null,
@@ -1758,7 +1843,7 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.10.8"
+   "version": "3.11.4"
   },
   "orig_nbformat": 4
  },
diff --git a/docs/torch/pytorch_transfomers.ipynb b/docs/torch/pytorch_transfomers.ipynb
new file mode 100644
index 0000000..e214cd1
--- /dev/null
+++ b/docs/torch/pytorch_transfomers.ipynb
@@ -0,0 +1,23 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "# Pytorch Transformer Implementation\n"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": []
+  }
+ ],
+ "metadata": {
+  "language_info": {
+   "name": "python"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 2
+}
diff --git a/requirements.txt b/requirements.txt
index 5f88124..1db4d1b 100644
--- a/requirements.txt
+++ b/requirements.txt
@@ -1,4 +1,4 @@
-Sphinx~=5.3
+Sphinx~=7.2
 -r docs/requirements.txt
 sphinx_rtd_theme
 sphinx-autobuild
\ No newline at end of file