What’s new in Llama 3

New Features

1) Parameters: It is trained with 8 billion and 70 billion parameters respectively. Therefore, a larger model with higher quantization is expected to outperform a smaller model with lower quantization.

2) Training Data: The token size has been increased to 15 trillion tokens, enabling more efficient language encoding and resulting in significantly enhanced model performance.

3) Context Length: The context length has also been extended from 4097 to 8192 tokens, providing a more extensive and detailed training corpus that enhances the model’s effectiveness.

4) Multi-lingual: Llama 3 now supports 30 languages, compared to Llama 2 which only supports 20 languages. This expansion allows for broader regional utilization of the model.

Llama 3 vs. LLaMa 2

1) Performance – Llama 3 8B contains 8 billion parameters and Llama 3 70B contains 70 billion parameters. Both can be considered big leaps in performance when compared to Llama 2 due to the training Llama 3 received using custom-built 24,000 GPU clusters.

2) Metrics – There are three important metrics that AI models aim to score high on. These are MMLU, ARC, and DROP. Llama 3 scores much higher than Llama 2 using these metrics, indicating a higher level of performance.

3) Data Training – Llama 3 is trained to encode with a wide range of languages.

4) Usability – Llama 3 has been fully integrated into Meta applications like Instagram and Facebook. Llama 2 was not available for use in this way, making it far less useable than the current version.

Here is a small overview of the difference between Llama3 and Llama2

 

How can we use LLAMA in Bagisto?

In Bagisto, we provide native support to various LLMs like Llama, OpenAI GPT-3.5-turbo, Mistral, Dolphin Phi, Phi-2, Starling, Llama 2 Uncensored, Llama 2 13B, Llama 2 70B, Orca Mini, Vicuna, LLaVA.

It provides Review Translation, Content Generation, and Personalized Checkout Messages.

Step 1 – Login to the Admin Panel of Bagisto and go to Configure >> Magic AI as shown below.

Step 2 – Now kindly Enable the Configurations for which you want to generate using Magic AI i.e. General Settings, Content Generation as shown in the below image.