From a254757a86e4129f4e84681965006ef3d353f750 Mon Sep 17 00:00:00 2001 From: Oscar Domingo Date: Fri, 23 Jul 2021 17:51:49 +0100 Subject: [PATCH] Update README.md This commit attempts to make a more streamlined experience when setting up DeepFaceLab on Linux. By being more explicit in the fact that you need to find out the correct `version for `cudnn` and `cudatoolkit`, as well as the added need of `-c conda-forge` for later versions of `cudatoolkit`. --- README.md | 29 ++++++++++++----------------- 1 file changed, 12 insertions(+), 17 deletions(-) diff --git a/README.md b/README.md index 8fe9f61bf..ac0b2b277 100644 --- a/README.md +++ b/README.md @@ -1,21 +1,17 @@ ## Using ### 1. Install Anaconda Anaconda is the preferred method of installing DeepFaceLab on Linux. -[Just follow the tutorial](https://docs.conda.io/projects/conda/en/latest/user-guide/install). +[Follow the Official tutorial](https://docs.conda.io/projects/conda/en/latest/user-guide/install). -### 2. Install System Dependencies -You will need FFMpeg, Git, and the most recent NVIDIA driver for your system to use this project. +### 2. Prerequisites +You need `FFMpeg`, `Git` and the most recent NVIDIA driver for your system. -If you are here, then you already have everything... - -### 3. Install DeepFaceLab - -Just run it in the terminal. - -Check latest cudnn and cudatoolkit version for your GPU device. +### 3. Install DeepFaceLab and DeepFaceLab for Linux +Check latest `cudnn` and `cudatoolkit` version for your GPU device, you can see your driver and CUDA version by issueing: +`nvidia-smi` and you can check in the [Official Compatibility List](https://docs.nvidia.com/deploy/cuda-compatibility/) to see which `cuDNN` and `cudatoolkit` version you should install. ```bash - conda create -n deepfacelab -c main python=3.7 cudnn=7.6.5 cudatoolkit=10.1.243 + conda create -n deepfacelab -c main -c conda-forge python=3.7 cudnn= cudatoolkit= conda activate deepfacelab git clone --depth 1 https://github.com/nagadit/DeepFaceLab_Linux.git cd DeepFaceLab_Linux @@ -23,10 +19,8 @@ Check latest cudnn and cudatoolkit version for your GPU device. python -m pip install -r ./DeepFaceLab/requirements-cuda.txt ``` -## 4. Download Pretrain (optional) -Use script 4.1 from the scripts directory. - -Or download manually +## 4. Optional: Download a Pre-trained model +Use any of the three scripts that start with `4.1_download_` to fetch them from the CLI, requires `wget` and `unzip`, alternatively you can download the manually: [CelebA](https://github.com/nagadit/DeepFaceLab_Linux/releases/download/1.0/pretrain_CelebA.zip) @@ -35,8 +29,9 @@ Or download manually [Quick96](https://github.com/nagadit/DeepFaceLab_Linux/releases/download/1.0/pretrain_Quick96.zip) ## 5. Navigate to the scripts directory and begin using DeepFaceLab_Linux ᗡ: -Run all scripts with BASH shell +Run all scripts with BASH shell, i.e: ```bash +cd scripts bash 1_clear_workspace.sh ``` -etc \ No newline at end of file +etc