Scribd
Upload
384 views

Installing Whisper WebUI On Windows 10 - Web

The document provides instructions for installing Whisper WebUI on Windows. This includes installing CUDA, Anaconda, Git, cloning the Whisper WebUI repository, creating a conda environment, installing PyTorch and other requirements, and running the WebUI.

Uploaded by

aadnk
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
384 views

Installing Whisper WebUI On Windows 10 - Web

The document provides instructions for installing Whisper WebUI on Windows. This includes installing CUDA, Anaconda, Git, cloning the Whisper WebUI repository, creating a conda environment, installing PyTorch and other requirements, and running the WebUI.

Uploaded by

aadnk
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 24

Installing Whisper WebUI on Windows

10/11
Install Nvidia CUDA Toolkit
First, go to https://developer.nvidia.com/cuda-downloads and select “Windows” next to “Operating
System”, “x86_64” next to “Architecture”, “10” or “11” next to “Version” and use the “exe
(Network)” installer Type:

Let the installer Launch:


Now, either leave “Express (Recommended)” selected, or choose “Custom (Advanced”):
If you chose “Custom (Advanced”), you may now customize the installation location. Then click
“Next”

Wait for the installer to finish (this may take 10-15 minutes depending on your CPU and network
speed):
Ignore “Nsight” if it pops up – just click “Next”:

The installer should then be finished – click “Close” (you may also want to deselect “Launch NVIDIA
GeForce Experience”)
Install Anaconda 3
Next, we will install the Python virtual environment manager Anaconda 3, to make managing the
Python environment easier and without any conflicts.

Go to https://www.anaconda.com/ and click “Download”:

Open the downloaded EXE:


Agree to the EULA:

You may then either install Anaconda for the current user or all users (requires admin):
Customize the installation folder:

Then start the installation process:


This may take a couple of minutes. Then click Next:
Git for Windows
Then, install Git for Windows. Go to https://gitforwindows.org/ and click Download:

And follow the installation wizard.


Checkout Whisper WebUI
Next, we will checkout the Whisper WebUI project to a folder. Navigate to a location where you’d
like to keep the Python project (for instance the Documents folder), and right click while holding the
SHIFT key:

Click “Open Powershell window here”. Then type the following command in the Powershell Window:

git clone "https://huggingface.co/spaces/aadnk/whisper-webui"


Close the window and navigate into the cloned folder:
Next, click the Windows button, and type “Anaconda Prompt” into the search window:
Click “Anaconda Prompt”. Then type the following command to create a new “whisper”
environment:
conda create --name whisper

Type “y” to create this environment:


Next, type “conda activate whisper” to actually use this environment. You will have to type
this every time you want to start the WebUI.

You should then check the Python version with the command “python --version”:

If this is different than Python 3.9.12, you may switch to Python 3.9.12 using the “conda install
python==3.9.12” command:
CondaSSLError
Now, for some reason, I encountered an error here – CondaSSLError: OpenSSL appears to be
unavailable on this machine.

This seems to be an issue with the Python version in the base package. A simple workaround is to
copy some DLL files as explained here:

• https://github.com/conda/conda/issues/11982#issuecomment-1285929629
You should then be able to install Python 3.9.12:

Type “y”.
Installing Torch

If you want GPU acceleration, you will also need to install the GPU version of PyTorch. Go to to the
following website:

• https://pytorch.org/get-started/locally/

Copy the command in “Run this command” and run it in the Anaconda Prompt window:
This may take some to execute:
Installing requirements

Start by copying the path to the folder “whisper-webui” from Explorer:

Execute the command “cd FOLDER-PATH”, where “FOLDER-PATH” is the path you copied from
Explorer:
Run the command “pip install -r requirements.txt”, which will install all other
requirements in addition to PyTorch:

This may take some time ...


Running WebUI

If you have closed “Anaconda Prompt” from the previous steps, or restarted your computer, you
may get back to this step by opening “Anaconda Prompt” by searching for it in Windows.
Then type “conda activate whisper” to load the Whisper environment. Finally type “cd FOLDER-
PATH” (where “FOLDER-PATH” is path to the whisper-webui directory you cloned earlier).

To run the UI, simply type the following command:


python app.py --input_audio_max_duration -1
Note that depending on the amount of VRAM on your GPU, you may only be able to run the “base”
or “small” model. You will get a “Out of memory” error if there’s not enough VRAM.

You might also like