<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:content="http://purl.org/rss/1.0/modules/content/"><channel><title>NLP on Ha Pham</title><link>http://hoanghapham.github.io/tags/nlp/</link><description>Recent content in NLP on Ha Pham</description><generator>Hugo -- 0.167.0</generator><language>en-us</language><lastBuildDate>Mon, 14 Jul 2025 16:07:05 +0700</lastBuildDate><atom:link href="http://hoanghapham.github.io/tags/nlp/index.xml" rel="self" type="application/rss+xml"/><item><title>Master's Thesis - HTR with Visual Language Model</title><link>http://hoanghapham.github.io/projects/vlm-htr/</link><pubDate>Mon, 14 Jul 2025 15:03:14 +0700</pubDate><guid>http://hoanghapham.github.io/projects/vlm-htr/</guid><description>&lt;p&gt;In this project, I fine-tuned Visual Language Models (VLMs) and create an end-to-end pipeline for Handwrittent Text Recognitition tasks on historical Swedish manuscripts. The VLM-based pipeline is also compared against the classical YOLO - TrOCR pipeline.&lt;/p&gt;
&lt;p&gt;The project was done in collaboration with the Swedish National Archives (Riksarkivet).&lt;/p&gt;
&lt;ul&gt;
&lt;li&gt;&lt;strong&gt;Models:&lt;/strong&gt; Florence-2, YOLO, TrOCR&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Tools:&lt;/strong&gt; PyTorch, transformers, Gradio&lt;/li&gt;
&lt;/ul&gt;
&lt;div class="github-link-wrapper"&gt;
&lt;a href="https://github.com/hoanghapham/vlm-htr" class="github-button" target="_blank" rel="noopener noreferrer"&gt;
&lt;i class="fab fa-github"&gt;&lt;/i&gt; GitHub
&lt;/a&gt;
&lt;/div&gt;
&lt;div class="generic-link-wrapper"&gt;
&lt;a href="https://huggingface.co/spaces/nazounoryuu/vlm-htr" class="generic-link" target="_blank" rel="noopener noreferrer"&gt;
&lt;i class="fa-solid fa-laptop-code"&gt;&lt;/i&gt; Demo
&lt;/a&gt;
&lt;/div&gt;</description></item><item><title>Language Detector</title><link>http://hoanghapham.github.io/projects/language-detector/</link><pubDate>Mon, 14 Jul 2025 16:07:05 +0700</pubDate><guid>http://hoanghapham.github.io/projects/language-detector/</guid><description>&lt;p&gt;In this project, I trained two models for the language detection task, using the WiLI-2018 dataset.&lt;/p&gt;
&lt;ul&gt;
&lt;li&gt;&lt;strong&gt;Models:&lt;/strong&gt; Naive Bayes, XLM RoBERTa&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Tools:&lt;/strong&gt; sklearn, transformers, Streamlit, FastAPI, Docker.&lt;/li&gt;
&lt;/ul&gt;
&lt;div class="github-link-wrapper"&gt;
&lt;a href="https://github.com/hoanghapham/language-detector" class="github-button" target="_blank" rel="noopener noreferrer"&gt;
&lt;i class="fab fa-github"&gt;&lt;/i&gt; GitHub
&lt;/a&gt;
&lt;/div&gt;</description></item></channel></rss>