Riley Learning

FLEURS Dataset: A Comprehensive Guide to Multilingual Speech Recognition

Introduction The FLEURS (Few-shot Learning Evaluation of Universal Representations of Speech) dataset represents a groundbreaking advancement in multilingual speech recognition research. Developed by Google, this comprehensive dataset provides high-quality speech data across 102 languages, making it an invaluable resource for researchers and practitioners working on multilingual automatic speech recognition (ASR)

Blog

Paper Review: Making LLMs Better Many-to-Many Speech-to-Text Translators with Curriculum Learning

In this post, I’m reviewing Making LLMs Better Many-to-Many Speech-to-Text Translators with Curriculum Learning (ACL 2025 ) without exotic training tricks. The recipe—called LLM-SRT—keeps a frozen Whisper encoder, adds a lightweight speech adapter (Q-Former + MLP), and fine-tunes in three stages: (1) ASR, (2) speech-aided MT (SMT/MMT), and

Blog

Paper Review - LongDocURL (ACL 2025)

Introduction Published at ACL 2025, "LongDocURL: a Comprehensive Multimodal Long Document Benchmark Integrating Understanding, Reasoning, and Locating" addresses important limitations in document understanding evaluation. This paper exposes fundamental gaps in how we test AI models on complex documents. The research, conducted by a collaborative team from Chinese Academy

Blog

Understanding and Comparing Embedding Models for RAG and Vector Search

In the rapidly evolving landscape of artificial intelligence, embedding models have emerged as the unsung heroes powering everything from semantic search to recommendation systems. These sophisticated models transform unstructured data into numerical vectors that preserve semantic meaning, enabling machines to understand and process human language with unprecedented accuracy. What Are

Blog

The Rise of On-Device AI: Comparing SLMs on NPUs and Copilot+ PCs

We're now witnessing the emergence of powerful on-device AI capabilities that promise to revolutionize how we interact with our computers. At the forefront of this transformation are SLMs running on NPUs, with Microsoft's Copilot+ PCs leading the charge into this new era.

Blog

Managing Python Projects: Virtual Environments, Git Ignore Rules, and Dependency Tracking

Effective Python project management is crucial for maintaining clean, reproducible, and collaborative codebases. This post covers the essential tools and practices for managing virtual environments, dependency tracking, and version control in Python projects.

Blog

Model Context Protocol (MCP): Shaping the Future of AI Agents

The Model Context Protocol (MCP) is an innovative protocol designed to enhance AI model interactions through advanced context management. This blog post explores what MCP is, how it works, and how developers can leverage its capabilities using the Python client example.

Blog

Understanding FastAPI: Building Production-Grade Asynchronous Applications with MCP

As the demand for real-time, responsive, and scalable AI applications grows, building robust asynchronous APIs becomes essential. In this guide, we explore FastAPI, a high-performance web framework for Python.

Blog

Building a Financial RAG Chatbot Using LLaMA, Streamlit and RunPod (VSCode)

In this tutorial, we’ll walk through the process of building a Financial Question-Answering chatbot using Retrieval-Augmented Generation (RAG), LLaMA 3, and Streamlit.

Blog

TensorFlow vs PyTorch vs Keras: A Beginner-Friendly Comparison of Deep Learning Frameworks

Whether you’re just stepping into the world of deep learning or already exploring complex neural networks, choosing the right framework is crucial. Among the many, three stand out: TensorFlow, PyTorch, and Keras.

RLHF vs. RLAIF: Fine-Tuning LLMs for Better Alignment (OTS, SFT, PPO, Jailbreak)

Large Language Models (LLMs) like GPT-4, LLaMA 3, and Claude are redefining natural language processing. Despite their advancements…

Optimizing Azure OpenAI Service: Base Model Deployment, Fine-Tuning, and Decoding Parameters

Azure OpenAI Service offers powerful tools to deploy, fine-tune, and interact with GPT models, making it essential to understand the…

RAG vs. Fine-Tuning : When to Use, Combine, and Optimize for Best Results

When building or optimizing AI models, two powerful techniques often come into play: Fine-tuning and RAG (Retrieval-Augmented Generation)…

Paper Review — Debug like a Human: A Large Language Model Debugger via Verifying Runtime Execution…

Debugging programs is essential yet challenging, even for advanced Large Language Models (LLMs). In their ACL 2024 paper, “Debug like a…

IMG2TEXT-Part2. OFA, CLIP Interrogator and ViT

Continuing from Part 1, we are going to look into the CLIP Interrogator, OFA model, and ViT model and ensemble them. Most of the codes are…

IMG2TEXT-Part1. Background (Stable Diffusion, CLIP, Prompt)

In this article, I’d like to talk about background information to implement CLIPInterrogator+OFA+ViT_LB0.568. Part 2 will cover the…

Google ISLR Transformer with W&B (Part 2)

In this article, I’ll be showing you how to create and train a model for the Kaggle ASL (American Sign Language) recognition competition…

Google ASL 1. Process Data with W&B 🐝

Today, I’m going to explain the dataset and how to process it for a Kaggle competition on ASL(American Sign Language), Google — Isolated…

Paper Review — Strided Transformer (TMM 2022)

Strided Transformer is a monocular 3D pose estimation model which lifts a long sequence of 2D joint locations to a single 3D pose.

Paper Review — VideoPose3D (CVPR 2019)

3D human pose estimation in video with temporal convolutions and semi-supervised training

Paper Review — MHFormer

Introduction

[PyTorch] Simple 3D Pose Baseline implementation (ICCV’17)

In this post, I review Simple 3D Pose Baseline (A simple yet effective baseline for 3d human pose estimation, also called as SIM) which is…

HRNet : Code Explained

HRNet(Deep High-Resolution Representation Learning for Human Pose Estimation) is a state-of-the-art algorithm in the field of semantic…

In this post, we create a simple convolutional neural network(SimpeConvNet) using only NumPy and…

Simple CovNet with NumPy In this post, we create a simple convolutional neural network(SimpeConvNet) using only NumPy and it will classify MNIST images. The codes are from a book called ‘Deep Learning from Scratch’. Let’s check an architecture of SimpeConvNet and notations first. Architecture N: the number of

Training Basic Two Layer Network with Numpy

In this post, we develop a two-layer network in order to perform classification in MNIST dataset and train it. There are mainly two parts…