From Release File (Windows)

Tested for Windows11

  • Make sure to have python (3.10 or 3.11) (Currently 3.12 is incompatible due to the version of numpy used) installed on your system and available in the PATH (There is a checkbox that you need to tick as soon as you run the installer to automatically do this). (If you have multiple version of python installed (including 3.12), make sure the first to appear in the PATH is a compatible version or set the PYTHON environment variable to the path of the compatible version in run-user.bat)

Python installer

  • From github download the release file

  • Unzip the release file downloaded in the previous step.

  • From inside the folder, run the run-user.bat file that will

    • Allow the user to tweak the environment variables to your liking (text editing the file), by setting/changing value to the already present ones, or adding new ones.

    • Create or reuse a virtual environment in the folder venv in the same directory as the script

    • Install the required packages to run the server

    • Start the server

The server will run with sensible defaults. Most notably the models files, plugin files (needed python packages) and database will be downloaded/created under %userprofile%/.ocr_translate. If not DEVICE is specified, the server will attempt to find the nvidia-smi executable and if available, will run on GPU.

For customization, you can set the Environment variables yourself.

Since the introduction of the run-user.bat script, this has been made easier as you just need to edit that file with any text editor and set the environment variables at the top of the file

@echo off

set LOAD_ON_START=

set DEVICE=cpu
set OCT_BASE_DIR=
set DATABASE_NAME=

set DJANGO_DEBUG=false
set DJANGO_LOG_LEVEL=INFO

call run.bat

You can see some of them already set to their default values or nothing, meaning they are left undefined.

The call run.bat calls the secondary script with the boilerplate code to start the server.