Installing Dependencies

Before installing dependencies

Make sure you have Microsoft Visual C++ 14.0 or greater installed To install it, go to this link

https://visualstudio.microsoft.com/visual-cpp-build-tools/

During install, select Desktop Development with C++, and then proceed to install

After install make sure to restart windows.

Setup/Dependencies

You will first need to setup the virtual environment using cuda The terminal command (on Windows) is:

python -m venv cuda

And then you will install your dependencies onto the virtual environment Use the following command to enter the virtual environment, when you are installing your dependencies or doing any other work afterwards:

cuda\Scripts\activate

You will need to install the following dependencies

1. matplotlib
2. numpy
3. pylzma
4. ipykernel
5. jupyter
6. torch

You can download all of the dependencies using this terminal command (on Windows):

pip3 install matplotlib numpy pylzma ipykernel jupyter

if you have errors with installing the dependencies, and the error tells you to update pip3 go ahead and use the command given by the error message to update pip3 in the virtual environment (Do not forget to go into the virtual environment first, using the command given above, before you do the update)

To install torch you can go to this link

https://pytorch.org/

Scroll down until you see this:

And choose your preferences and use the link given to you at the bottom For reference, the image contains my preferences and the command I used

pip3 install torch torchvision --index-url https://download.pytorch.org/whl/cu128

Things/Commands to Know

Remember all commands and work is to be done in the python virtual environment Refer to README line 30 to find the command to use the virtual environment that was setup

Once all the dependencies have been installed, you can open the jupyter notebook using this command:

jupyter notebook

You need to make sure you're running the jupyter code in cuda, so we create a kernel and on the jupyter notebook you change the kernel to whatever the display name you set. For reference, this is the command I used to initialize my cuda kernel:

python -m ipykernel install --user --name=cuda --display-name "MerlinLLM"

Then to change the kernel you are working on, in any of the .ipynb files When you run the command on README line 78, you will be directed to this page

Then click on the .ipynb file you're going to work on and it will take you here Then select the Kernel tab and select Change Kernel

You will get a screen like this:

Just go ahead and choose the kernel that corresponds to the display name you created (Mine is MerlinLLM)

If you are looking for information on all the different torch functions (at least the relevant ones), then you can refer to the playground.ipynb file that contains code that shows how each function works, what the output looks like, and relevant comments on how to use the functions.