Skip to main content

Deep Experiments

Search This Blog

Deep Learning: Experimenting Neural Style Transfer on Floyd Cloud

Get link
Facebook
X
Pinterest
Email
Other Apps

Get link
Facebook
X
Pinterest
Email
Other Apps

Comments

Post a Comment

Popular Posts

FAISS Index: Your magic wand to find things in a High-dimensional mess (OOPs Maze!) Imagine searching for a specific face in a crowd of a million people. Sounds daunting, right? But for computers dealing with vast collections of data points, this can be simplified to a similarity-based search . That's where FAISS comes in, a lightning-fast indexing library specially designed for finding similar items within high-dimensional spaces. Why do we need FAISS? Well, data nowadays exists in complex, multi-dimensional forms. We have facial images represented as thousands of numbers, music described by intricate frequency patterns, and even documents encoded into numerical vectors. Finding similar examples within these massive datasets using brute-force comparisons is simply impractical. FAISS steps in, compressing and organizing these vectors into efficient structures, allowing rapid retrieval of similar ones. FAISS index types: Flat indexes : Simple vector encod

How to try open-source LLMs locally?

Running open-source LLMs locally has benefits like data privacy and cost savings, but it requires some technical know-how. Here's a breakdown to get you started: Before you begin: Hardware considerations: Since general-purpose LLMs are resource-intensive, your computer should have enough RAM (ideally 16GB or more) and a powerful graphics card (GPU) for optimal performance. My experiments are with i9 16GB laptop with NVIDIA 4070, with Ubuntu 22.04 LTS as OS. Approaches to try LLMs locally: All-in-one Desktop Solutions: These are ideal for beginners. Tools like GPT4All offer user-friendly interfaces that mimic popular LLM services like ChatGPT. You simply download and launch the application, and it handles the technical setup for you. Command Line Interface (CLI) & Backend Servers: This approach offers more flexibility but requires some technical knowledge. There are many LLM frameworks like GPT4All, LM Studio, Jan, llama.cpp, llamafile, Ollama, and NextChat, that do

Contrastive Language-Image Pre-training (CLIP): What is to know?

What is CLIP Paradigm in multi-modal machine learning? Quickly re-purposing a painstaking trained computer vision (CV) model for a new task is often a desired goal while optimizing the engineering pipelines for practical applications. However effectiveness of learning image representations was inherently limited by the training set of images and new data required substantial retraining. Contrastive Language-Image Pre-training (CLIP) is a method of learning image representation and associated text representations together . Training: CLIP jointly trains an image encoder and a text encoder. During the training, model is fed image+text pairs and is encouraged to predict which of the possible (image, text) pairings across a batch actually occurred. Thus CLIP ends up learning a multi-modal embedding space by jointly training an image encoder and text encoder to maximise the cosine similarity of the image and text embedding of its description, and at the same time minimisin

Powered by Blogger