In this article, we summarize the Metal Llama 3 technical report with the most crucial aspects such as the architecture, the pretraining data, the compute infrastructure, the post-training strategy, and multimodal capabilities of Llama 3. ...

In this article, we create a multimodal RAG application from scratch to chat with PDFs, text files, images, and videos using the Phi-3.5 family of language models. ...

In this article, we create a multimodal chat interface with Gradio to chat with images and videos using Phi-3.5 Vision Instruct model. ...

In this article, we use LitGPT, LitAPI, and LitServe for serving LLMs using Lightning Studio and also on the local system. ...

In this article, we use different Torchvision backbones for creating DeepLab segmentation models and train it on the Pascal VOC semantic segmentation dataset. ...