April 2026
Intermediate
395 pages
13h 22m
English
Like LLMs, LMMs are advanced AI systems. But in contrast to LLMs, LMMs are capable of understanding and possibly creating multiple types (modalities) of data, such as text, images, audio, and even video. Figure 1.12 shows possible input and output modalities of an LMM.
Figure 1.12 Large Multimodal Models
These models integrate diverse modalities into a single model. These diverse modalities can be either on the input and/or the output of the model. Some models allow for both text and images as inputs and provide text as output, called vision-language models (VLMs). Popular examples are Llama 3.2 Vision, GPT-4o, ...
Read now
Unlock full access