Install LLAMA 2 locally
In this article, we will show step by step on how to set up LLAMA 2 in a local environment of your choice. It will also be necessary to examine the assessment of the model on different cases to evaluate its effectiveness.

Step 1: Install Python 3.10.9
First of all, you need to install Python 3.10. It is recommended to install the ninth version of Python on your computer which is available on the Python official website. Here it is suggested to use this particular version, because at the moment there is an incompatibility with Python 3.11 and the Torch library.
Step 2: Install Miniconda
Next, you are required to download and install Miniconda on your system To do this, follow the ensuing methodology: The Windows installer 64-bit can be downloaded from the Miniconda homepage. For more information visit Miniconda Documentation.
Step 3: Create a New Conda Environment
Once you have installed Miniconda, open the Anaconda prompt and create a new environment using the following command:

Step 4: Activate the Environment
To activate the environment, use the following command:

Step 5: Install Required Libraries
Next, you need to install the required libraries using pip:

Step 6: Install Text Generation Web UI
To set up the web UI for text generation, execute the command provided below:
Replace username with the actual username of the repository owner.

Step 7: Install Requirements
Move over to the cloned repository and install the prerequisites using the following command:

Step 8: Run the Server
Finally, start the server by run the following command:

When you have made sure that the server is running, you can open your web browser and type http://localhost:7860 to obtain the text generation web user interface.

Testing the Model
We tested the LLaMA 2 model on various use cases, including:
- Writing a Python script to output numbers 1 to 100
- Writing a snake game in Python
- Writing a poem about AI in exactly 50 words
- Writing an email to a boss letting them know you are leaving the company
- Answering factual questions
- Testing math skills
For the majority of the tasks set, the model was quite effective, but it failed to respond correctly to a query that it would have provided data regarding unlawful acts.