From c494be58cfbf44b5882a02ec54814a52da01ec41 Mon Sep 17 00:00:00 2001
From: Sotiris Gkouzias <82544018+sgkouzias@users.noreply.github.com>
Date: Mon, 8 Apr 2024 14:20:57 +0300
Subject: [PATCH 01/15] Update pip.md

Specify additional steps to utilize GPU for Linux users
---
 site/en/install/pip.md | 62 ++++++++++++++++++++++++++++++++++++++++--
 1 file changed, 60 insertions(+), 2 deletions(-)

diff --git a/site/en/install/pip.md b/site/en/install/pip.md
index 4add60b11d7..c6168e0536f 100644
--- a/site/en/install/pip.md
+++ b/site/en/install/pip.md
@@ -193,7 +193,35 @@ The following NVIDIA® software are only required for GPU support.
     nvidia-smi
     ```
 
-    ### 3. Install TensorFlow
+    ### 3. Install Miniconda
+
+    [Miniconda](https://docs.conda.io/en/latest/miniconda.html){:.external}
+    is the recommended approach for installing TensorFlow with GPU support.
+    It creates a separate environment to avoid changing any installed
+    software in your system. This is also the easiest way to install the
+    required software especially for the GPU setup.
+
+    Follow the instuctions of the conda user guide to install miniconda
+    [Miniconda Installation Guide](https://conda.io/projects/conda/en/latest/user-guide/install/linux.html){:.external}.
+
+    ### 4. Create a conda environment
+
+    Create a new conda environment named `tf` with the following command.
+
+    ```bash
+    conda create --name tf python=3.11
+    ```
+
+    You can deactivate and activate it with the following commands.
+
+    ```bash
+    conda deactivate
+    conda activate tf
+    ```
+
+    Make sure it is activated for the rest of the installation.
+    
+    ### 5. Install TensorFlow
 
     TensorFlow requires a recent version of pip, so upgrade your pip
     installation to be sure you're running the latest version.
@@ -211,7 +239,37 @@ The following NVIDIA® software are only required for GPU support.
     pip install tensorflow
     ```
 
-    ### 4. Verify the installation
+    ### 6. Set environment variables
+    Locate the directory for the conda environment in your terminal window by running in the terminal:
+    `echo $CONDA_PREFIX`
+
+    Enter that directory and create these subdirectories and files:
+
+    ```bash
+    cd $CONDA_PREFIX
+    mkdir -p ./etc/conda/activate.d
+    mkdir -p ./etc/conda/deactivate.d
+    touch ./etc/conda/activate.d/env_vars.sh
+    touch ./etc/conda/deactivate.d/env_vars.sh
+    ```
+    Edit `./etc/conda/activate.d/env_vars.sh` as follows:
+
+    ```bash
+    #!/bin/sh
+
+    export NVIDIA_DIR=$(dirname $(dirname $(python -c "import nvidia.cudnn;print(nvidia.cudnn.__file__)")))
+    export LD_LIBRARY_PATH=$(echo ${NVIDIA_DIR}/*/lib/ | sed -r 's/\s+/:/g')${LD_LIBRARY_PATH:+:${LD_LIBRARY_PATH}}
+    ```
+    Edit `./etc/conda/deactivate.d/env_vars.sh` as follows:
+
+    ```bash
+    #!/bin/sh
+
+    unset NVIDIA_DIR
+    unset LD_LIBRARY_PATH
+    ```
+
+    ### 7. Verify the installation
 
     Verify the CPU setup:
 

From 40824e44a075ccd1d5f54f2fd5500ceeae871297 Mon Sep 17 00:00:00 2001
From: Sotiris Gkouzias <82544018+sgkouzias@users.noreply.github.com>
Date: Mon, 8 Apr 2024 22:44:48 +0300
Subject: [PATCH 02/15] Update pip.md

Advice to skip additional step 6 if using CPU.
---
 site/en/install/pip.md | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/site/en/install/pip.md b/site/en/install/pip.md
index c6168e0536f..c1d86f2573a 100644
--- a/site/en/install/pip.md
+++ b/site/en/install/pip.md
@@ -240,6 +240,9 @@ The following NVIDIA® software are only required for GPU support.
     ```
 
     ### 6. Set environment variables
+
+    You can skip this section if you only run TensorFlow on the CPU.
+
     Locate the directory for the conda environment in your terminal window by running in the terminal:
     `echo $CONDA_PREFIX`
 

From 840fec9c3183a5555e24feee14ddd63cd182991a Mon Sep 17 00:00:00 2001
From: Sotiris Gkouzias <82544018+sgkouzias@users.noreply.github.com>
Date: Tue, 9 Apr 2024 21:58:38 +0300
Subject: [PATCH 03/15] Update pip.md

Added second option to create virtual env via Python's built in venv module for Linux users with CUDA-enabled GPUs
---
 site/en/install/pip.md | 44 +++++++++++++++++++++++++++++++++++++-----
 1 file changed, 39 insertions(+), 5 deletions(-)

diff --git a/site/en/install/pip.md b/site/en/install/pip.md
index c1d86f2573a..93cae82503a 100644
--- a/site/en/install/pip.md
+++ b/site/en/install/pip.md
@@ -195,6 +195,8 @@ The following NVIDIA® software are only required for GPU support.
 
     ### 3. Install Miniconda
 
+    You can skip this section if you prefer to use Python’s built in venv module instead of Miniconda.
+
     [Miniconda](https://docs.conda.io/en/latest/miniconda.html){:.external}
     is the recommended approach for installing TensorFlow with GPU support.
     It creates a separate environment to avoid changing any installed
@@ -204,22 +206,30 @@ The following NVIDIA® software are only required for GPU support.
     Follow the instuctions of the conda user guide to install miniconda
     [Miniconda Installation Guide](https://conda.io/projects/conda/en/latest/user-guide/install/linux.html){:.external}.
 
-    ### 4. Create a conda environment
+    ### 4. Create a virtual environment
+
+    #### First option: Miniconda 
 
     Create a new conda environment named `tf` with the following command.
 
     ```bash
     conda create --name tf python=3.11
     ```
+    #### Second option: [venv](https://docs.python.org/3/library/venv.html){:.external}
 
-    You can deactivate and activate it with the following commands.
+    Navigate to your desired virtual environments' directory and create a new venv environment named `tf` with the following command.
 
     ```bash
-    conda deactivate
-    conda activate tf
+    python3 -m venv tf 
     ```
 
-    Make sure it is activated for the rest of the installation.
+    You can activate it with the following command.
+
+    ```bash
+    source tf/bin/activate
+    ```
+
+    Make sure that the virtual environment you just created is activated for the rest of the installation.
     
     ### 5. Install TensorFlow
 
@@ -243,6 +253,8 @@ The following NVIDIA® software are only required for GPU support.
 
     You can skip this section if you only run TensorFlow on the CPU.
 
+    #### First option: Miniconda 
+
     Locate the directory for the conda environment in your terminal window by running in the terminal:
     `echo $CONDA_PREFIX`
 
@@ -271,6 +283,28 @@ The following NVIDIA® software are only required for GPU support.
     unset NVIDIA_DIR
     unset LD_LIBRARY_PATH
     ```
+    #### Second option: venv
+
+    Locate the directory for the venv environment in your terminal window by running in the terminal:
+    `echo $VIRTUAL_ENV`
+
+    Enter that directory and add the following lines at the end of the activate script `./bin/activate` as follows:
+    
+    ```bash
+    # Set environment variables for TensorFlow
+    export NVIDIA_DIR=$(dirname $(dirname $(python -c "import nvidia.cudnn;print(nvidia.cudnn.__file__)")))
+    export LD_LIBRARY_PATH=$(echo ${NVIDIA_DIR}/*/lib/ | sed -r 's/\s+/:/g')${LD_LIBRARY_PATH:+:${LD_LIBRARY_PATH}}
+    ```
+    
+    Add the following lines at the end of `deactivate` block:
+    
+    ```bash
+    deactivate () {
+    # ...
+    unset NVIDIA_DIR
+    unset LD_LIBRARY_PATH
+    }
+    ```
 
     ### 7. Verify the installation
 

From 5448363ce65bcec0c0320a4fb2add79edd3f3f76 Mon Sep 17 00:00:00 2001
From: Sotiris Gkouzias <82544018+sgkouzias@users.noreply.github.com>
Date: Wed, 10 Apr 2024 08:17:34 +0300
Subject: [PATCH 04/15] Update pip.md

Added virtual envs activation/deactivation commands and changed wording for editing the deactivate block in the activate script of the venv virtual env.
---
 site/en/install/pip.md | 14 +++++++++++---
 1 file changed, 11 insertions(+), 3 deletions(-)

diff --git a/site/en/install/pip.md b/site/en/install/pip.md
index 93cae82503a..27c9e278ab4 100644
--- a/site/en/install/pip.md
+++ b/site/en/install/pip.md
@@ -215,6 +215,13 @@ The following NVIDIA® software are only required for GPU support.
     ```bash
     conda create --name tf python=3.11
     ```
+    You can activate and deactivate it with the following commands.
+
+    ```bash
+    conda activate tf
+    conda deactivate
+    ```
+
     #### Second option: [venv](https://docs.python.org/3/library/venv.html){:.external}
 
     Navigate to your desired virtual environments' directory and create a new venv environment named `tf` with the following command.
@@ -223,13 +230,14 @@ The following NVIDIA® software are only required for GPU support.
     python3 -m venv tf 
     ```
 
-    You can activate it with the following command.
+    You can activate and deactivate it with the following commands.
 
     ```bash
     source tf/bin/activate
+    deactivate
     ```
 
-    Make sure that the virtual environment you just created is activated for the rest of the installation.
+    Make sure that the virtual environment is activated for the rest of the installation.
     
     ### 5. Install TensorFlow
 
@@ -296,7 +304,7 @@ The following NVIDIA® software are only required for GPU support.
     export LD_LIBRARY_PATH=$(echo ${NVIDIA_DIR}/*/lib/ | sed -r 's/\s+/:/g')${LD_LIBRARY_PATH:+:${LD_LIBRARY_PATH}}
     ```
     
-    Add the following lines at the end of `deactivate` block:
+    Add the following lines at the end of `deactivate` block in the activate script to ensure that the necessary CUDA environment variables are set only while the virtual environment is active:
     
     ```bash
     deactivate () {

From c7518a2d0e0099c28aed404117bf202e8ca453c4 Mon Sep 17 00:00:00 2001
From: Sotiris Gkouzias <82544018+sgkouzias@users.noreply.github.com>
Date: Wed, 10 Apr 2024 22:07:20 +0300
Subject: [PATCH 05/15] Update pip.md

Added instructions to resolve the ptxas issue.
---
 site/en/install/pip.md | 58 +++++++++++++++++++++++++++++++++++-------
 1 file changed, 49 insertions(+), 9 deletions(-)

diff --git a/site/en/install/pip.md b/site/en/install/pip.md
index 27c9e278ab4..b62f341088f 100644
--- a/site/en/install/pip.md
+++ b/site/en/install/pip.md
@@ -280,16 +280,32 @@ The following NVIDIA® software are only required for GPU support.
     ```bash
     #!/bin/sh
 
-    export NVIDIA_DIR=$(dirname $(dirname $(python -c "import nvidia.cudnn;print(nvidia.cudnn.__file__)")))
-    export LD_LIBRARY_PATH=$(echo ${NVIDIA_DIR}/*/lib/ | sed -r 's/\s+/:/g')${LD_LIBRARY_PATH:+:${LD_LIBRARY_PATH}}
+    # Store original LD_LIBRARY_PATH 
+    export ORIGINAL_LD_LIBRARY_PATH="${LD_LIBRARY_PATH}" 
+
+    # Get the directory of CUDA libraries
+    CUDA_DIR=$(dirname $(dirname $(python -c "import nvidia.cuda_nvcc; print(nvidia.cuda_nvcc.__file__)")))
+
+    # Set LD_LIBRARY_PATH to include all directories containing CUDA libraries
+    export LD_LIBRARY_PATH=$(find ${CUDA_DIR}/*/lib/ -type d -printf "%p:")${LD_LIBRARY_PATH:+:${LD_LIBRARY_PATH}}
+
+    # Get the directory of NVCC binaries
+    NVCC_DIR=$(dirname $(dirname $(python -c "import nvidia.cuda_nvcc; print(nvidia.cuda_nvcc.__file__)")))
+
+    # Set PATH to include all directories containing NVCC binaries
+    export PATH=$(find ${NVCC_DIR}/*/bin/ -type d -printf "%p:")${PATH:+:${PATH}}
     ```
     Edit `./etc/conda/deactivate.d/env_vars.sh` as follows:
 
     ```bash
     #!/bin/sh
+    
+    # Restore original LD_LIBRARY_PATH
+    export LD_LIBRARY_PATH="${ORIGINAL_LD_LIBRARY_PATH}"
 
-    unset NVIDIA_DIR
-    unset LD_LIBRARY_PATH
+    # Unset environment variables
+    unset CUDA_DIR
+    unset NVCC_DIR
     ```
     #### Second option: venv
 
@@ -299,9 +315,20 @@ The following NVIDIA® software are only required for GPU support.
     Enter that directory and add the following lines at the end of the activate script `./bin/activate` as follows:
     
     ```bash
-    # Set environment variables for TensorFlow
-    export NVIDIA_DIR=$(dirname $(dirname $(python -c "import nvidia.cudnn;print(nvidia.cudnn.__file__)")))
-    export LD_LIBRARY_PATH=$(echo ${NVIDIA_DIR}/*/lib/ | sed -r 's/\s+/:/g')${LD_LIBRARY_PATH:+:${LD_LIBRARY_PATH}}
+    # Store the original LD_LIBRARY_PATH
+    export ORIGINAL_LD_LIBRARY_PATH=$LD_LIBRARY_PATH 
+
+    # Get the directory of CUDA libraries
+    CUDA_DIR=$(dirname $(dirname $(python -c "import nvidia.cuda_nvcc; print(nvidia.cuda_nvcc.__file__)")))
+    
+    # Set LD_LIBRARY_PATH to include all directories containing CUDA libraries
+    export LD_LIBRARY_PATH=$(find ${CUDA_DIR}/*/lib/ -type d -printf "%p:")${LD_LIBRARY_PATH:+:${LD_LIBRARY_PATH}}
+
+    # Get the directory of NVCC binaries
+    NVCC_DIR=$(dirname $(dirname $(python -c "import nvidia.cuda_nvcc; print(nvidia.cuda_nvcc.__file__)")))
+    
+    # Set PATH to include all directories containing NVCC binaries
+    export PATH=$(find ${NVCC_DIR}/*/bin/ -type d -printf "%p:")${PATH:+:${PATH}}
     ```
     
     Add the following lines at the end of `deactivate` block in the activate script to ensure that the necessary CUDA environment variables are set only while the virtual environment is active:
@@ -309,8 +336,21 @@ The following NVIDIA® software are only required for GPU support.
     ```bash
     deactivate () {
     # ...
-    unset NVIDIA_DIR
-    unset LD_LIBRARY_PATH
+    # Unset the added path to PATH if within a virtual environment
+    if [ -n "$VIRTUAL_ENV" ]; then
+        # Remove the path from PATH
+        PATH=$(echo $PATH | sed -e "s|${NVCC_DIR}/*/bin/:||g")
+    fi
+
+    # Restore the original LD_LIBRARY_PATH
+    if [ -n "$ORIGINAL_LD_LIBRARY_PATH" ]; then
+        export LD_LIBRARY_PATH=$ORIGINAL_LD_LIBRARY_PATH
+        unset ORIGINAL_LD_LIBRARY_PATH
+    fi
+
+    # Unset environment variables
+    unset CUDA_DIR
+    unset NVCC_DIR    
     }
     ```
 

From 6a40fe47c629737803bbdbb171e5b4c45fd93e57 Mon Sep 17 00:00:00 2001
From: Sotiris Gkouzias <82544018+sgkouzias@users.noreply.github.com>
Date: Wed, 10 Apr 2024 23:28:32 +0300
Subject: [PATCH 06/15] Update pip.md

Revised CUDNN_DIR definition
---
 site/en/install/pip.md | 54 +++++++++++++++++++++---------------------
 1 file changed, 27 insertions(+), 27 deletions(-)

diff --git a/site/en/install/pip.md b/site/en/install/pip.md
index b62f341088f..75f2229a002 100644
--- a/site/en/install/pip.md
+++ b/site/en/install/pip.md
@@ -283,10 +283,10 @@ The following NVIDIA® software are only required for GPU support.
     # Store original LD_LIBRARY_PATH 
     export ORIGINAL_LD_LIBRARY_PATH="${LD_LIBRARY_PATH}" 
 
-    # Get the directory of CUDA libraries
-    CUDA_DIR=$(dirname $(dirname $(python -c "import nvidia.cuda_nvcc; print(nvidia.cuda_nvcc.__file__)")))
+    # Get the CUDNN directory 
+    CUDNN_DIR=$(dirname $(dirname $(python -c "import nvidia.cudnn; print(nvidia.cudnn.__file__)")))
 
-    # Set LD_LIBRARY_PATH to include all directories containing CUDA libraries
+    # Set LD_LIBRARY_PATH to include CUDNN directory
     export LD_LIBRARY_PATH=$(find ${CUDA_DIR}/*/lib/ -type d -printf "%p:")${LD_LIBRARY_PATH:+:${LD_LIBRARY_PATH}}
 
     # Get the directory of NVCC binaries
@@ -304,7 +304,7 @@ The following NVIDIA® software are only required for GPU support.
     export LD_LIBRARY_PATH="${ORIGINAL_LD_LIBRARY_PATH}"
 
     # Unset environment variables
-    unset CUDA_DIR
+    unset CUDNN_DIR
     unset NVCC_DIR
     ```
     #### Second option: venv
@@ -315,14 +315,14 @@ The following NVIDIA® software are only required for GPU support.
     Enter that directory and add the following lines at the end of the activate script `./bin/activate` as follows:
     
     ```bash
-    # Store the original LD_LIBRARY_PATH
+    # Store original LD_LIBRARY_PATH
     export ORIGINAL_LD_LIBRARY_PATH=$LD_LIBRARY_PATH 
 
-    # Get the directory of CUDA libraries
-    CUDA_DIR=$(dirname $(dirname $(python -c "import nvidia.cuda_nvcc; print(nvidia.cuda_nvcc.__file__)")))
+    # Get the CUDNN directory
+    CUDNN_DIR=$(dirname $(dirname $(python -c "import nvidia.cudnn; print(nvidia.cudnn.__file__)")))
     
-    # Set LD_LIBRARY_PATH to include all directories containing CUDA libraries
-    export LD_LIBRARY_PATH=$(find ${CUDA_DIR}/*/lib/ -type d -printf "%p:")${LD_LIBRARY_PATH:+:${LD_LIBRARY_PATH}}
+    # Set LD_LIBRARY_PATH to include CUDNN directory
+    export LD_LIBRARY_PATH=$(find ${CUDNN_DIR}/*/lib/ -type d -printf "%p:")${LD_LIBRARY_PATH:+:${LD_LIBRARY_PATH}}
 
     # Get the directory of NVCC binaries
     NVCC_DIR=$(dirname $(dirname $(python -c "import nvidia.cuda_nvcc; print(nvidia.cuda_nvcc.__file__)")))
@@ -331,26 +331,26 @@ The following NVIDIA® software are only required for GPU support.
     export PATH=$(find ${NVCC_DIR}/*/bin/ -type d -printf "%p:")${PATH:+:${PATH}}
     ```
     
-    Add the following lines at the end of `deactivate` block in the activate script to ensure that the necessary CUDA environment variables are set only while the virtual environment is active:
-    
+    Add the following lines at the end of `deactivate` block in the activate script to ensure that the necessary NVIDIA environment variables are set only while the virtual environment is active:
+
     ```bash
     deactivate () {
-    # ...
-    # Unset the added path to PATH if within a virtual environment
-    if [ -n "$VIRTUAL_ENV" ]; then
-        # Remove the path from PATH
-        PATH=$(echo $PATH | sed -e "s|${NVCC_DIR}/*/bin/:||g")
-    fi
-
-    # Restore the original LD_LIBRARY_PATH
-    if [ -n "$ORIGINAL_LD_LIBRARY_PATH" ]; then
-        export LD_LIBRARY_PATH=$ORIGINAL_LD_LIBRARY_PATH
-        unset ORIGINAL_LD_LIBRARY_PATH
-    fi
-
-    # Unset environment variables
-    unset CUDA_DIR
-    unset NVCC_DIR    
+        # ...
+        # Unset the added path to PATH if within a virtual environment
+        if [ -n "$VIRTUAL_ENV" ]; then
+            # Remove the path from PATH
+            PATH=$(echo $PATH | sed -e "s|${NVCC_DIR}/*/bin/:||g")
+        fi
+
+        # Restore original LD_LIBRARY_PATH
+        if [ -n "$ORIGINAL_LD_LIBRARY_PATH" ]; then
+            export LD_LIBRARY_PATH=$ORIGINAL_LD_LIBRARY_PATH
+            unset ORIGINAL_LD_LIBRARY_PATH
+        fi
+
+        # Unset environment variables
+        unset CUDNN_DIR
+        unset NVCC_DIR    
     }
     ```
 

From 82713bffccf5c16293cd70b140729d4a77e2a8a5 Mon Sep 17 00:00:00 2001
From: Sotiris Gkouzias <82544018+sgkouzias@users.noreply.github.com>
Date: Wed, 10 Apr 2024 23:45:13 +0300
Subject: [PATCH 07/15] Update pip.md

Corrected LD_LIBRARY_PATH definition in conda environment instructions
---
 site/en/install/pip.md | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/site/en/install/pip.md b/site/en/install/pip.md
index 75f2229a002..7c4a63bb9dd 100644
--- a/site/en/install/pip.md
+++ b/site/en/install/pip.md
@@ -287,7 +287,7 @@ The following NVIDIA® software are only required for GPU support.
     CUDNN_DIR=$(dirname $(dirname $(python -c "import nvidia.cudnn; print(nvidia.cudnn.__file__)")))
 
     # Set LD_LIBRARY_PATH to include CUDNN directory
-    export LD_LIBRARY_PATH=$(find ${CUDA_DIR}/*/lib/ -type d -printf "%p:")${LD_LIBRARY_PATH:+:${LD_LIBRARY_PATH}}
+    export LD_LIBRARY_PATH=$(find ${CUDNN_DIR}/*/lib/ -type d -printf "%p:")${LD_LIBRARY_PATH:+:${LD_LIBRARY_PATH}}
 
     # Get the directory of NVCC binaries
     NVCC_DIR=$(dirname $(dirname $(python -c "import nvidia.cuda_nvcc; print(nvidia.cuda_nvcc.__file__)")))

From b81e4f2eb80eb29911cc2723c9b2bdcaa3460c20 Mon Sep 17 00:00:00 2001
From: Sotiris Gkouzias <82544018+sgkouzias@users.noreply.github.com>
Date: Fri, 12 Apr 2024 10:04:18 +0300
Subject: [PATCH 08/15] Update pip.md

Rename environment variable to PTXAS_DIR and package manager options.
---
 site/en/install/pip.md | 40 +++++++++++++++++++++-------------------
 1 file changed, 21 insertions(+), 19 deletions(-)

diff --git a/site/en/install/pip.md b/site/en/install/pip.md
index 7c4a63bb9dd..572a081430a 100644
--- a/site/en/install/pip.md
+++ b/site/en/install/pip.md
@@ -195,7 +195,7 @@ The following NVIDIA® software are only required for GPU support.
 
     ### 3. Install Miniconda
 
-    You can skip this section if you prefer to use Python’s built in venv module instead of Miniconda.
+    You can skip this section if you have already installed `Miniconda` (referred as *option #1* in the next steps) or you prefer to use Python’s built-in `venv` module (referred as *option #2* in the next steps) instead.
 
     [Miniconda](https://docs.conda.io/en/latest/miniconda.html){:.external}
     is the recommended approach for installing TensorFlow with GPU support.
@@ -208,7 +208,7 @@ The following NVIDIA® software are only required for GPU support.
 
     ### 4. Create a virtual environment
 
-    #### First option: Miniconda 
+    * ***Option #1: Miniconda*** 
 
     Create a new conda environment named `tf` with the following command.
 
@@ -222,9 +222,11 @@ The following NVIDIA® software are only required for GPU support.
     conda deactivate
     ```
 
-    #### Second option: [venv](https://docs.python.org/3/library/venv.html){:.external}
-
-    Navigate to your desired virtual environments' directory and create a new venv environment named `tf` with the following command.
+    * ***Option #2: venv***
+      
+    The [venv](https://docs.python.org/3/library/venv.html){:.external} module supports creating lightweight “virtual environments”, each with their own independent set of Python packages installed in their site directories. 
+    
+    Navigate to your desired virtual environments directory and create a new venv environment named `tf` with the following command.
 
     ```bash
     python3 -m venv tf 
@@ -261,7 +263,7 @@ The following NVIDIA® software are only required for GPU support.
 
     You can skip this section if you only run TensorFlow on the CPU.
 
-    #### First option: Miniconda 
+    * ***Option #1: Miniconda*** 
 
     Locate the directory for the conda environment in your terminal window by running in the terminal:
     `echo $CONDA_PREFIX`
@@ -289,11 +291,11 @@ The following NVIDIA® software are only required for GPU support.
     # Set LD_LIBRARY_PATH to include CUDNN directory
     export LD_LIBRARY_PATH=$(find ${CUDNN_DIR}/*/lib/ -type d -printf "%p:")${LD_LIBRARY_PATH:+:${LD_LIBRARY_PATH}}
 
-    # Get the directory of NVCC binaries
-    NVCC_DIR=$(dirname $(dirname $(python -c "import nvidia.cuda_nvcc; print(nvidia.cuda_nvcc.__file__)")))
+    # Get the ptxas directory  
+    PTXAS_DIR=$(dirname $(dirname $(python -c "import nvidia.cuda_nvcc; print(nvidia.cuda_nvcc.__file__)")))
 
-    # Set PATH to include all directories containing NVCC binaries
-    export PATH=$(find ${NVCC_DIR}/*/bin/ -type d -printf "%p:")${PATH:+:${PATH}}
+    # Set PATH to include the directory containing ptxas
+    export PATH=$(find ${PTXAS_DIR}/*/bin/ -type d -printf "%p:")${PATH:+:${PATH}}
     ```
     Edit `./etc/conda/deactivate.d/env_vars.sh` as follows:
 
@@ -305,9 +307,9 @@ The following NVIDIA® software are only required for GPU support.
 
     # Unset environment variables
     unset CUDNN_DIR
-    unset NVCC_DIR
+    unset PTXAS_DIR
     ```
-    #### Second option: venv
+    * ***Option #2: venv***
 
     Locate the directory for the venv environment in your terminal window by running in the terminal:
     `echo $VIRTUAL_ENV`
@@ -324,11 +326,11 @@ The following NVIDIA® software are only required for GPU support.
     # Set LD_LIBRARY_PATH to include CUDNN directory
     export LD_LIBRARY_PATH=$(find ${CUDNN_DIR}/*/lib/ -type d -printf "%p:")${LD_LIBRARY_PATH:+:${LD_LIBRARY_PATH}}
 
-    # Get the directory of NVCC binaries
-    NVCC_DIR=$(dirname $(dirname $(python -c "import nvidia.cuda_nvcc; print(nvidia.cuda_nvcc.__file__)")))
+    # Get the ptxas directory 
+    PTXAS_DIR=$(dirname $(dirname $(python -c "import nvidia.cuda_nvcc; print(nvidia.cuda_nvcc.__file__)")))
     
-    # Set PATH to include all directories containing NVCC binaries
-    export PATH=$(find ${NVCC_DIR}/*/bin/ -type d -printf "%p:")${PATH:+:${PATH}}
+    # Set PATH to include the directory containing ptxas
+    export PATH=$(find ${PTXAS_DIR}/*/bin/ -type d -printf "%p:")${PATH:+:${PATH}}
     ```
     
     Add the following lines at the end of `deactivate` block in the activate script to ensure that the necessary NVIDIA environment variables are set only while the virtual environment is active:
@@ -336,10 +338,10 @@ The following NVIDIA® software are only required for GPU support.
     ```bash
     deactivate () {
         # ...
-        # Unset the added path to PATH if within a virtual environment
+        # Unset the added path to PATH if within the virtual environment
         if [ -n "$VIRTUAL_ENV" ]; then
             # Remove the path from PATH
-            PATH=$(echo $PATH | sed -e "s|${NVCC_DIR}/*/bin/:||g")
+            PATH=$(echo $PATH | sed -e "s|${PTXAS_DIR}/*/bin/:||g")
         fi
 
         # Restore original LD_LIBRARY_PATH
@@ -350,7 +352,7 @@ The following NVIDIA® software are only required for GPU support.
 
         # Unset environment variables
         unset CUDNN_DIR
-        unset NVCC_DIR    
+        unset PTXAS_DIR    
     }
     ```
 

From aebf305d3401691d83a44179bdbf60363012cf49 Mon Sep 17 00:00:00 2001
From: Sotiris Gkouzias <82544018+sgkouzias@users.noreply.github.com>
Date: Fri, 12 Apr 2024 10:13:49 +0300
Subject: [PATCH 09/15] Update pip.md

Added note to use pip instead of conda to install TensorFlow.
---
 site/en/install/pip.md | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/site/en/install/pip.md b/site/en/install/pip.md
index 572a081430a..4b428467732 100644
--- a/site/en/install/pip.md
+++ b/site/en/install/pip.md
@@ -259,6 +259,8 @@ The following NVIDIA® software are only required for GPU support.
     pip install tensorflow
     ```
 
+    Note: Do not install TensorFlow with `conda`. It may not have the latest stable version. `pip` is recommended since TensorFlow is only officially released to PyPI.
+
     ### 6. Set environment variables
 
     You can skip this section if you only run TensorFlow on the CPU.

From 146bbebfdd8138f2317781761162b928871721be Mon Sep 17 00:00:00 2001
From: Sotiris Gkouzias <82544018+sgkouzias@users.noreply.github.com>
Date: Mon, 17 Jun 2024 22:28:15 +0300
Subject: [PATCH 10/15] Simplify procedure by removing option to install with
 conda virtual env.

Removed option to install within conda virtual environment. Recommendation to install in venv environment.
---
 site/en/install/pip.md | 87 +++---------------------------------------
 1 file changed, 5 insertions(+), 82 deletions(-)

diff --git a/site/en/install/pip.md b/site/en/install/pip.md
index 4b428467732..ce0cd95503f 100644
--- a/site/en/install/pip.md
+++ b/site/en/install/pip.md
@@ -193,38 +193,9 @@ The following NVIDIA® software are only required for GPU support.
     nvidia-smi
     ```
 
-    ### 3. Install Miniconda
-
-    You can skip this section if you have already installed `Miniconda` (referred as *option #1* in the next steps) or you prefer to use Python’s built-in `venv` module (referred as *option #2* in the next steps) instead.
-
-    [Miniconda](https://docs.conda.io/en/latest/miniconda.html){:.external}
-    is the recommended approach for installing TensorFlow with GPU support.
-    It creates a separate environment to avoid changing any installed
-    software in your system. This is also the easiest way to install the
-    required software especially for the GPU setup.
-
-    Follow the instuctions of the conda user guide to install miniconda
-    [Miniconda Installation Guide](https://conda.io/projects/conda/en/latest/user-guide/install/linux.html){:.external}.
-
-    ### 4. Create a virtual environment
-
-    * ***Option #1: Miniconda*** 
-
-    Create a new conda environment named `tf` with the following command.
-
-    ```bash
-    conda create --name tf python=3.11
-    ```
-    You can activate and deactivate it with the following commands.
-
-    ```bash
-    conda activate tf
-    conda deactivate
-    ```
-
-    * ***Option #2: venv***
+    ### 3. Create a virtual environment with [venv](https://docs.python.org/3/library/venv.html){:.external}
       
-    The [venv](https://docs.python.org/3/library/venv.html){:.external} module supports creating lightweight “virtual environments”, each with their own independent set of Python packages installed in their site directories. 
+    The venv module is part of Python’s standard library and is the officially recommended way to create virtual environments. 
     
     Navigate to your desired virtual environments directory and create a new venv environment named `tf` with the following command.
 
@@ -241,7 +212,7 @@ The following NVIDIA® software are only required for GPU support.
 
     Make sure that the virtual environment is activated for the rest of the installation.
     
-    ### 5. Install TensorFlow
+    ### 4. Install TensorFlow
 
     TensorFlow requires a recent version of pip, so upgrade your pip
     installation to be sure you're running the latest version.
@@ -261,58 +232,10 @@ The following NVIDIA® software are only required for GPU support.
 
     Note: Do not install TensorFlow with `conda`. It may not have the latest stable version. `pip` is recommended since TensorFlow is only officially released to PyPI.
 
-    ### 6. Set environment variables
+    ### 5. Set environment variables
 
     You can skip this section if you only run TensorFlow on the CPU.
 
-    * ***Option #1: Miniconda*** 
-
-    Locate the directory for the conda environment in your terminal window by running in the terminal:
-    `echo $CONDA_PREFIX`
-
-    Enter that directory and create these subdirectories and files:
-
-    ```bash
-    cd $CONDA_PREFIX
-    mkdir -p ./etc/conda/activate.d
-    mkdir -p ./etc/conda/deactivate.d
-    touch ./etc/conda/activate.d/env_vars.sh
-    touch ./etc/conda/deactivate.d/env_vars.sh
-    ```
-    Edit `./etc/conda/activate.d/env_vars.sh` as follows:
-
-    ```bash
-    #!/bin/sh
-
-    # Store original LD_LIBRARY_PATH 
-    export ORIGINAL_LD_LIBRARY_PATH="${LD_LIBRARY_PATH}" 
-
-    # Get the CUDNN directory 
-    CUDNN_DIR=$(dirname $(dirname $(python -c "import nvidia.cudnn; print(nvidia.cudnn.__file__)")))
-
-    # Set LD_LIBRARY_PATH to include CUDNN directory
-    export LD_LIBRARY_PATH=$(find ${CUDNN_DIR}/*/lib/ -type d -printf "%p:")${LD_LIBRARY_PATH:+:${LD_LIBRARY_PATH}}
-
-    # Get the ptxas directory  
-    PTXAS_DIR=$(dirname $(dirname $(python -c "import nvidia.cuda_nvcc; print(nvidia.cuda_nvcc.__file__)")))
-
-    # Set PATH to include the directory containing ptxas
-    export PATH=$(find ${PTXAS_DIR}/*/bin/ -type d -printf "%p:")${PATH:+:${PATH}}
-    ```
-    Edit `./etc/conda/deactivate.d/env_vars.sh` as follows:
-
-    ```bash
-    #!/bin/sh
-    
-    # Restore original LD_LIBRARY_PATH
-    export LD_LIBRARY_PATH="${ORIGINAL_LD_LIBRARY_PATH}"
-
-    # Unset environment variables
-    unset CUDNN_DIR
-    unset PTXAS_DIR
-    ```
-    * ***Option #2: venv***
-
     Locate the directory for the venv environment in your terminal window by running in the terminal:
     `echo $VIRTUAL_ENV`
 
@@ -358,7 +281,7 @@ The following NVIDIA® software are only required for GPU support.
     }
     ```
 
-    ### 7. Verify the installation
+    ### 6. Verify the installation
 
     Verify the CPU setup:
 

From 7cf1c57de806b1527e53385ff5af0d56f5bd4caa Mon Sep 17 00:00:00 2001
From: Sotiris Gkouzias <82544018+sgkouzias@users.noreply.github.com>
Date: Tue, 18 Jun 2024 19:09:02 +0300
Subject: [PATCH 11/15] Instructions to create symlinks to NVIDIA shared
 libraries and ptxas.

Replaced instructions to modify default activate/deactivate scripts with instructions to create symlinks to NVIDIA shared libraries and ptxas.
---
 site/en/install/pip.md | 46 ++++++++----------------------------------
 1 file changed, 8 insertions(+), 38 deletions(-)

diff --git a/site/en/install/pip.md b/site/en/install/pip.md
index ce0cd95503f..2b4558a523c 100644
--- a/site/en/install/pip.md
+++ b/site/en/install/pip.md
@@ -232,53 +232,23 @@ The following NVIDIA® software are only required for GPU support.
 
     Note: Do not install TensorFlow with `conda`. It may not have the latest stable version. `pip` is recommended since TensorFlow is only officially released to PyPI.
 
-    ### 5. Set environment variables
+    ### 5. Virtual environment configuration
 
     You can skip this section if you only run TensorFlow on the CPU.
 
-    Locate the directory for the venv environment in your terminal window by running in the terminal:
-    `echo $VIRTUAL_ENV`
-
-    Enter that directory and add the following lines at the end of the activate script `./bin/activate` as follows:
+    * Create symbolic links to NVIDIA shared libraries:
     
     ```bash
-    # Store original LD_LIBRARY_PATH
-    export ORIGINAL_LD_LIBRARY_PATH=$LD_LIBRARY_PATH 
-
-    # Get the CUDNN directory
-    CUDNN_DIR=$(dirname $(dirname $(python -c "import nvidia.cudnn; print(nvidia.cudnn.__file__)")))
-    
-    # Set LD_LIBRARY_PATH to include CUDNN directory
-    export LD_LIBRARY_PATH=$(find ${CUDNN_DIR}/*/lib/ -type d -printf "%p:")${LD_LIBRARY_PATH:+:${LD_LIBRARY_PATH}}
-
-    # Get the ptxas directory 
-    PTXAS_DIR=$(dirname $(dirname $(python -c "import nvidia.cuda_nvcc; print(nvidia.cuda_nvcc.__file__)")))
-    
-    # Set PATH to include the directory containing ptxas
-    export PATH=$(find ${PTXAS_DIR}/*/bin/ -type d -printf "%p:")${PATH:+:${PATH}}
+    pushd $(dirname $(python -c 'print(__import__("tensorflow").__file__)'))
+    ln -svf ../nvidia/*/lib/*.so* .
+    popd
     ```
     
-    Add the following lines at the end of `deactivate` block in the activate script to ensure that the necessary NVIDIA environment variables are set only while the virtual environment is active:
+    * Create a symbolic link to ptxas:
 
     ```bash
-    deactivate () {
-        # ...
-        # Unset the added path to PATH if within the virtual environment
-        if [ -n "$VIRTUAL_ENV" ]; then
-            # Remove the path from PATH
-            PATH=$(echo $PATH | sed -e "s|${PTXAS_DIR}/*/bin/:||g")
-        fi
-
-        # Restore original LD_LIBRARY_PATH
-        if [ -n "$ORIGINAL_LD_LIBRARY_PATH" ]; then
-            export LD_LIBRARY_PATH=$ORIGINAL_LD_LIBRARY_PATH
-            unset ORIGINAL_LD_LIBRARY_PATH
-        fi
-
-        # Unset environment variables
-        unset CUDNN_DIR
-        unset PTXAS_DIR    
-    }
+    ln -sf $(find $(dirname $(dirname $(python -c "import nvidia.cuda_nvcc;         
+    print(nvidia.cuda_nvcc.__file__)"))/*/bin/) -name ptxas -print -quit) $VIRTUAL_ENV/bin/ptxas
     ```
 
     ### 6. Verify the installation

From 3046cdc2390e13acf1817661072c61e0dfa0b5b2 Mon Sep 17 00:00:00 2001
From: Sotiris Gkouzias <82544018+sgkouzias@users.noreply.github.com>
Date: Wed, 19 Jun 2024 21:19:17 +0300
Subject: [PATCH 12/15] Removed deactivate command.

---
 site/en/install/pip.md | 5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)

diff --git a/site/en/install/pip.md b/site/en/install/pip.md
index 2b4558a523c..19d7b6469a2 100644
--- a/site/en/install/pip.md
+++ b/site/en/install/pip.md
@@ -203,11 +203,10 @@ The following NVIDIA® software are only required for GPU support.
     python3 -m venv tf 
     ```
 
-    You can activate and deactivate it with the following commands.
+    You can activate it with the following command.
 
     ```bash
-    source tf/bin/activate
-    deactivate
+    source tf/bin/activate    
     ```
 
     Make sure that the virtual environment is activated for the rest of the installation.

From 7f5cce6e97b740b7d0dcb79edcd00c54d3db6ec2 Mon Sep 17 00:00:00 2001
From: Sotiris Gkouzias <82544018+sgkouzias@users.noreply.github.com>
Date: Tue, 2 Jul 2024 19:42:14 +0300
Subject: [PATCH 13/15] Instructions to create symlinks in case the intended
 way doesn't work.

Revised the step with instructions to configure the virtual environment variables for GPU users by adding a disclaimer.
---
 site/en/install/pip.md | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)

diff --git a/site/en/install/pip.md b/site/en/install/pip.md
index 19d7b6469a2..28e5360e38f 100644
--- a/site/en/install/pip.md
+++ b/site/en/install/pip.md
@@ -229,7 +229,7 @@ The following NVIDIA® software are only required for GPU support.
     pip install tensorflow
     ```
 
-    Note: Do not install TensorFlow with `conda`. It may not have the latest stable version. `pip` is recommended since TensorFlow is only officially released to PyPI.
+    **Note:** Do not install TensorFlow with `conda`. It may not have the latest stable version. `pip` is recommended since TensorFlow is only officially released to PyPI.
 
     ### 5. Virtual environment configuration
 
@@ -249,6 +249,10 @@ The following NVIDIA® software are only required for GPU support.
     ln -sf $(find $(dirname $(dirname $(python -c "import nvidia.cuda_nvcc;         
     print(nvidia.cuda_nvcc.__file__)"))/*/bin/) -name ptxas -print -quit) $VIRTUAL_ENV/bin/ptxas
     ```
+    *️⃣ **Disclaimer**
+
+    Symbolic links are only necessary in case the intended way doesn't work, i.e. the components aren't being
+    detected, and/or conflict with the existing system CUDA installation.
 
     ### 6. Verify the installation
 

From 64e7c5097ad570ac825ff4d3e203e81e6020dcd2 Mon Sep 17 00:00:00 2001
From: Sotiris Gkouzias <82544018+sgkouzias@users.noreply.github.com>
Date: Wed, 3 Jul 2024 09:01:17 +0300
Subject: [PATCH 14/15] Reformat disclaimer on environment configuration step.

Deleted asterisk emoji and placed disclaimer note before respective instructions.
---
 site/en/install/pip.md | 9 ++++-----
 1 file changed, 4 insertions(+), 5 deletions(-)

diff --git a/site/en/install/pip.md b/site/en/install/pip.md
index 28e5360e38f..f96d743e395 100644
--- a/site/en/install/pip.md
+++ b/site/en/install/pip.md
@@ -234,6 +234,9 @@ The following NVIDIA® software are only required for GPU support.
     ### 5. Virtual environment configuration
 
     You can skip this section if you only run TensorFlow on the CPU.
+    
+    Note: Symbolic links are only necessary in case the intended way doesn't work, i.e. the components aren't being
+    detected, and/or conflict with the existing system CUDA installation.
 
     * Create symbolic links to NVIDIA shared libraries:
     
@@ -249,11 +252,7 @@ The following NVIDIA® software are only required for GPU support.
     ln -sf $(find $(dirname $(dirname $(python -c "import nvidia.cuda_nvcc;         
     print(nvidia.cuda_nvcc.__file__)"))/*/bin/) -name ptxas -print -quit) $VIRTUAL_ENV/bin/ptxas
     ```
-    *️⃣ **Disclaimer**
-
-    Symbolic links are only necessary in case the intended way doesn't work, i.e. the components aren't being
-    detected, and/or conflict with the existing system CUDA installation.
-
+    
     ### 6. Verify the installation
 
     Verify the CPU setup:

From 06798ba3737d595e8c5a7880b0f3fdc63053d0ff Mon Sep 17 00:00:00 2001
From: Mark Daoust <markdaoust@google.com>
Date: Wed, 4 Sep 2024 14:38:58 -0700
Subject: [PATCH 15/15] Move the GPU debugging step to after the "Test the
 installation" step.

---
 site/en/install/pip.md | 41 ++++++++++++++++++++++++-----------------
 1 file changed, 24 insertions(+), 17 deletions(-)

diff --git a/site/en/install/pip.md b/site/en/install/pip.md
index f96d743e395..670f54b874e 100644
--- a/site/en/install/pip.md
+++ b/site/en/install/pip.md
@@ -231,12 +231,29 @@ The following NVIDIA® software are only required for GPU support.
 
     **Note:** Do not install TensorFlow with `conda`. It may not have the latest stable version. `pip` is recommended since TensorFlow is only officially released to PyPI.
 
-    ### 5. Virtual environment configuration
+    ### 6. Verify the installation
 
-    You can skip this section if you only run TensorFlow on the CPU.
-    
-    Note: Symbolic links are only necessary in case the intended way doesn't work, i.e. the components aren't being
-    detected, and/or conflict with the existing system CUDA installation.
+    Verify the CPU setup:
+
+    ```bash
+    python3 -c "import tensorflow as tf; print(tf.reduce_sum(tf.random.normal([1000, 1000])))"
+    ```
+
+    If a tensor is returned, you've installed TensorFlow successfully.
+
+    Verify the GPU setup:
+
+    ```bash
+    python3 -c "import tensorflow as tf; print(tf.config.list_physical_devices('GPU'))"
+    ```
+
+    If a list of GPU devices is returned, you've installed TensorFlow
+    successfully. **If not continue to the next step**.
+
+    ### 6. [GPU only] Virtual environment configuration
+
+    If the GPU test in the last section was unsuccessful, the most likely cause is that components aren't being detected,
+    and/or conflict with the existing system CUDA installation. So you need to add some symbolic links to fix this.
 
     * Create symbolic links to NVIDIA shared libraries:
     
@@ -252,16 +269,6 @@ The following NVIDIA® software are only required for GPU support.
     ln -sf $(find $(dirname $(dirname $(python -c "import nvidia.cuda_nvcc;         
     print(nvidia.cuda_nvcc.__file__)"))/*/bin/) -name ptxas -print -quit) $VIRTUAL_ENV/bin/ptxas
     ```
-    
-    ### 6. Verify the installation
-
-    Verify the CPU setup:
-
-    ```bash
-    python3 -c "import tensorflow as tf; print(tf.reduce_sum(tf.random.normal([1000, 1000])))"
-    ```
-
-    If a tensor is returned, you've installed TensorFlow successfully.
 
     Verify the GPU setup:
 
@@ -269,8 +276,8 @@ The following NVIDIA® software are only required for GPU support.
     python3 -c "import tensorflow as tf; print(tf.config.list_physical_devices('GPU'))"
     ```
 
-    If a list of GPU devices is returned, you've installed TensorFlow
-    successfully.
+
+
 
 *   {MacOS}