| ... | ... | @@ -9,96 +9,39 @@ This new version will be available on **October 17th** ! |
|
|
|
|
|
|
|
----
|
|
|
|
|
|
|
|
**- Prerequisites -**
|
|
|
|
## Prerequisites
|
|
|
|
|
|
|
|
* All notebooks can run on a **classic CPU**, having a GPU can be nice, but is not essential.
|
|
|
|
* Original datasets require about 2.5G, but Consolidated datasets may require up to 150G depending on the tests you perform. **15GB** will probably be a minimum...
|
|
|
|
|
|
|
|
**- Important : -**
|
|
|
|
|
|
|
|
- If you know **how to use Docker tool** :<br>
|
|
|
|
## 1 - Installation via Docker (Best choice !)
|
|
|
|
|
|
|
|
If you know **how to use Docker tool** :<br>
|
|
|
|
You can [**install the Fidle environment via Docker**](Install-Docker) !
|
|
|
|
([<img width="15px" src="uploads/677b49fe6e5f0aee65658ba58afb94ce/noun-star-2324309-E12229.png">
|
|
|
|
<img width="15px" src="uploads/677b49fe6e5f0aee65658ba58afb94ce/noun-star-2324309-E12229.png">
|
|
|
|
<img width="15px" src="uploads/677b49fe6e5f0aee65658ba58afb94ce/noun-star-2324309-E12229.png">](Install-Docker))
|
|
|
|
<br> The Docker image contains everything you need.
|
|
|
|
<br> This is objectively the best choice !
|
|
|
|
|
|
|
|
## 2 - Conventional installation
|
|
|
|
|
|
|
|
- **Otherwise** :
|
|
|
|
This page explains how to install Fidle environment "the classical way".
|
|
|
|
|
|
|
|
**- Contents -**
|
|
|
|
|
|
|
|
[[_TOC_]]
|
|
|
|
|
|
|
|
|
|
|
|
## Step 0 - Create a main repository
|
|
|
|
|
|
|
|
The very first step will be to create a folder in which we will then install all the elements of Fidle, for example : `fidle-tp`
|
|
|
|
|
|
|
|
The next 3 steps will take place within this folder.
|
|
|
|
|
|
|
|
## Step 1 - Get Fidle repository of notebooks
|
|
|
|
|
|
|
|
The simplest and most conventional way is with **git** (200 MB):
|
|
|
|
|
|
|
|
```shell
|
|
|
|
git clone https://gricad-gitlab.univ-grenoble-alpes.fr/talks/fidle.git
|
|
|
|
```
|
|
|
|
|
|
|
|
However, a **[classic retrieval](https://fidle.cnrs.fr/notebooks)** (zip) is possible via the download button, near \[Clone\] button.
|
|
|
|
|
|
|
|
**Important :** The project is frequently updated and we encourage you to get the latest version.\
|
|
|
|
The version number is specified in the README of the repository.
|
|
|
|
|
|
|
|
## Step 2 - Get and install Datasets
|
|
|
|
|
|
|
|
**Get it**\
|
|
|
|
The datasets used in the examples are **[available here ](https://fidle.cnrs.fr/fidle-datasets.tar)**.\
|
|
|
|
On Linux, you can retrieve and decompress them using the following commands:
|
|
|
|
|
|
|
|
```shell
|
|
|
|
wget https://fidle.cnrs.fr/fidle-datasets.tar
|
|
|
|
tar -xf fidle-datasets.tar
|
|
|
|
```
|
|
|
|
|
|
|
|
A complete description of the different datasets is available in Notebooks.
|
|
|
|
|
|
|
|
## Step 3 - Setup Python Environment
|
|
|
|
This procedure will allow you to install all the environment necessary for the Fidle practical work. This environment is identical in every way to a classic Deep Learning development environment.
|
|
|
|
|
|
|
|
Choose the right procedure :
|
|
|
|
|
|
|
|
- [Linux installation](/Using-Fidle/Linux installation with pip)
|
|
|
|
- [Windows installation](/Using Fidle/Windows installation)
|
|
|
|
- [Manual installation (MacOs ?)](/Using Fidle/Manual installation)
|
|
|
|
|
|
|
|
If you want to **check your tensorflow installation** (and check if it supports gpu), launch a python interpreter and use these instructions :
|
|
|
|
|
|
|
|
```python
|
|
|
|
import tensorflow as tf
|
|
|
|
print(tf.__version__)
|
|
|
|
print(tf.test.is_built_with_cuda())
|
|
|
|
print(tf.config.list_physical_devices('GPU'))
|
|
|
|
```
|
|
|
|
|
|
|
|
Great, now you can **[run Jupyter Lab !](Using-Fidle/Running-Jupyter)**
|
|
|
|
|
|
|
|
|
|
|
|
## Annexe 1 : Hide warning messages
|
|
|
|
|
|
|
|
- To avoid infos and warning messages from tensorflow, you can set an additional environment variable `TF_CPP_MIN_LOG_LEVEL` whose possible values are 0, 1, 2 or 3.\
|
|
|
|
(0: all msg, 1:hide infos, 2:hide infos+warnings, 3:hide info, warnings and errors)
|
|
|
|
|
|
|
|
- **[Linux installation](/Using-Fidle/Linux installation with pip)**
|
|
|
|
- **[Windows installation](/Using Fidle/Windows installation)**
|
|
|
|
- **[Manual installation (MacOs ?)](/Using Fidle/Manual installation)**
|
|
|
|
|
|
|
|
## Annexe 2 - Uninstall
|
|
|
|
Then, now you can **[run Jupyter Lab !](Using-Fidle/Running-Jupyter)**
|
|
|
|
|
|
|
|
To remove and delete all Fidle stuff :
|
|
|
|
- delete the folder containing the datasets (probably `fidle-datasets`)
|
|
|
|
- delete the folder containing the notebooks (probably `fidle` ou `fidle-master`)
|
|
|
|
- close terminals and remove python virtual env directory (probably `file-env`)
|
|
|
|
|
|
|
|
...and think to remove your FIDLE_DATASET_DIR variable setting !
|
|
|
|
## 3 - Uninstall
|
|
|
|
|
|
|
|
To remove and delete all Fidle stuff : Just remove your Fidle folder
|
|
|
|
(`fidle-tp` in our examples)
|
|
|
|
|
|
|
|
|
|
|
|
|
| ... | ... | |
