# 深度學習python套件安裝 (with Ubuntu22.04)
###### tags: `Tutorial`, `Ubuntu`, `python`

目錄：
+ [Pytorch](#pytorch)
+ [Jax](#jax)
+ [Tensorflow](#tf)
+ [Pyspark](#pyspark)
+ [Kaggle API](#kaggle)

其他參考：
+ [Ubuntu22.04重灌筆記](https://hackmd.io/@Chieh997/UbuntuCUDAvenv)

---

<a id='pytorch'></a>
## Pytorch
### [Installation](https://pytorch.org/get-started/locally/#start-locally)

pip install torch torchvision torchaudio

### Cuda check

import torch
torch.cuda.is_available()
torch.cuda.device_count()
torch.cuda.current_device()
torch.cuda.device(0)
torch.cuda.get_device_name(0)

<a id='jax'></a>
## Jax
### [Installation](https://github.com/jax-ml/jax)
+ CPU only (Linux/MacOS/Windows)

pip install -U jax

+ GPU (NVIDIA with CUDA 12, Linux)

pip install -U "jax[cuda12]"

::: warning
(2025.02.18)
This version (via pip) doesn't support for CUDNN 9. (XlaRuntimeError: INTERNAL: the library was not initialized)
To use the latest CUDNN, try [install locally](https://docs.jax.dev/en/latest/installation.html#pip-installation-nvidia-gpu-cuda-installed-locally-harder) instead.
:::

### Problems
+ Jax(<=0.5.0) was built with NumPy 1.25.0, doesn't support for NumPy>=2.0. Downgrade it to specific version is needed.

pip install --upgrade numpy==1.25.0

<a id='tf'></a>
## TensorFlow
### [Installation](https://www.tensorflow.org/install/pip?hl=zh-tw)*
`Tensorflow2` 最新穩定版本，支援 CPU 和 GPU (Ubuntu 和 Windows)，無須額外指定。

pip install --upgrade tensorflow

但目前版本(2.11)可能與TensorRT 8.x不相容，建議參考下面的安裝方式。

### Problems
#### Could not load dynamic library 'libnvinfer.so.7'
+ [Could not load dynamic library 'libnvinfer.so.7'](https://stackoverflow.com/questions/74956134/could-not-load-dynamic-library-libnvinfer-so-7)
+ [TensorFlow 2.10.0 not compatible with TensorRT 8.4.3](https://github.com/tensorflow/tensorflow/issues/57679#issuecomment-1249197802)

先安裝新版TensorRT(8.x)

pip install --upgrade setuptools pip
pip install nvidia-pyindex
pip install nvidia-tensorrt
# verify installation of tensorrt
python3 -c "import tensorrt; print(tensorrt.__version__); assert tensorrt.Builder(tensorrt.Logger())"

建立`libnvinfer` version 7 和 8 的 symbolic link

# 到tensorrt的安裝位置(因安裝方式不同有異)
cd /env/lib/python3.10/site-packages/tensorrt
# create symbolic links
ln -s libnvinfer_plugin.so.8 libnvinfer_plugin.so.7
ln -s libnvinfer.so.8 libnvinfer.so.7

更改.bashrc或venv/activate，新增`tensorrt`至`LD_LIBRARY_PATH`

export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:~/env/lib/python3.10/site-packages/tensorrt/

安裝Tensorflow並驗證GPU安裝是否成功

pip install --upgrade tensorflow
python3 -c "import tensorflow as tf; print(tf.config.list_physical_devices('GPU'))"

#### Fixing "successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node"
+ [What is NUMA (Non-Uniformed Memory Access)](https://gist.github.com/zrruziev/b93e1292bf2ee39284f834ec7397ee9f)

1. Check Nodes

lspci | grep -i nvidia

輸出結果如下：

01:00.0 VGA compatible controller: NVIDIA Corporation TU106 [GeForce RTX {xxxx}] (rev a1)
01:00.1 Audio device: NVIDIA Corporation TU106 High Definition Audio Controller (rev a1)

表示 VGA compatible device, NVIDIA Geforce 在 `01:00`，如果顯示結果不同可能須視情況調整下列code。

2. Check if it is connected.

cat /sys/bus/pci/devices/0000\:01\:00.0/numa_node

+ `0`表示已連結。`-1`表示未連結，須執行下一步驟。

3. Fix it with the command below.

sudo echo 0 | sudo tee -a /sys/bus/pci/devices/0000\:01\:00.0/numa_node

可再次執行步驟2確認是否連結成功。

#### Memory limit setting
+ [Setting GPUs memory limit for each notebook.](https://stackoverflow.com/questions/69401285/attempting-to-perform-blas-operation-using-streamexecutor-without-blas-support)
> 僅建議作為替代方案，最好還是能順利安裝TensorRT，不然還是會有很多bug。

如果TensorRT安裝失敗/沒有成功抓到，可能在使用Jupyter Notebook時會遇到GPU記憶體不足的問題，此時可以通過設定每個Notebook的可用記憶體限制來解決。

import tensorflow as tf
gpus = tf.config.experimental.list_physical_devices('GPU')
if gpus:
  try:
    tf.config.experimental.set_virtual_device_configuration(
        gpus[0],[tf.config.experimental.VirtualDeviceConfiguration(memory_limit=4096)])
  except RuntimeError as e:
    print(e)

<a id='pyspark'></a>
## Pyspark
### Install Pyspark on Ubuntu
+ [To install Apache Spark and run Pyspark in Ubuntu 22.04](https://dev.to/kinyungu_denis/to-install-apache-spark-and-run-pyspark-in-ubuntu-2204-4i79)
+ [Installing Apache Spark on Ubuntu](https://linuxhint.com/install-apache-spark-ubuntu/)
+ [Installation of PySpark (All operating systems)](https://www.datacamp.com/tutorial/installation-of-pyspark)
### [Use Pyspark on Jupyter Notebook and Virtualenv](https://www.sicara.fr/blog-technique/2017-05-02-get-started-pyspark-jupyter-notebook-3-minutes)

進入虛擬環境，安裝`findpyspark`

pip install findspark

在Notebook的起始或python script的起始使用下列代碼：

import findspark
findspark.init()

<a id='kaggle'></a>
## [Kaggle API](https://github.com/Kaggle/kaggle-api)
進入虛擬環境，安裝`kaggle`

pip install kaggle --upgrade

### API credentials
要使用API下載競賽或其他資料，則必須要先設定你的API token才能正常使用。

1. 在官網注冊帳號，再到帳號的介面中選取'Create API Token'即會下載一個`kaggle.json`的檔案，當中包含你的API token資料。
2. 設定API token，設定的方法可以為下列任一種：
+ 將Token放在指定路徑`$HOME/.kaggle/kaggle.json`下，並運行下列指令防止其他電腦使用者讀取你的API

chmod 600 ~/.kaggle/kaggle.json

+ 將username和key (參考`kaggle.json`) export到虛擬環境中

export KAGGLE_USERNAME={user_name}
export KAGGLE_KEY={xxxxxxxxxxxxxx}