ML Application · Jupyter Notebook

doc-assistant-hf

An AI-powered Document Intelligence tool built with Hugging Face Transformers and Gradio. Extracts precise answers and generates summaries from PDF, DOCX, and TXT files using specialized local models.

Pop Alexandru@pop123-ux · Atlas 70.7 · #4= builder

View on GitHub2 Open Pop Alexandru's Atlas

What it is

The repository provides a document intelligence notebook that fine-tunes a BART summarizer with QLoRA, runs extractive QA using a Longformer model with custom span decoding, and implements a TextRank-based extractive summarizer, all wrapped in a Gradio interface.

Scored from the code at its last scan, 8 Oct 2026. How merit is scored

Where it counts

NLP & LLMs85%
Machine Learning65%
Data Science20%
MLOps10%

How much of the project is about each field, as scoring counts it. It ranks in the field it is most about.

Around it on the NLP & LLMs board

Whole board

RankProjectBuilderFieldMerit

25apexLibrary / Framework · RustThe repository provides the reference Rust implementation of the APEX (Agent Planning & Execution Specification) DSL, including a parser, validator, interpreter, and tool registry for controlling LLM reasoning and agent workflows.54.8/100ImplementationLuiz Cesar SpiesNLP & LLMs
26qsbcArchitecture Modification / Research · Jupyter NotebookThe repository provides infrastructure for the Quantized Semantic Bottleneck Classifier (QSBC) architecture applied to the C3PA privacy-policy dataset, including a data pipeline, agent for generating reasoning, web explorer, and ablation experiments to validate the architecture.54.1/100ImplementationJarad PeckenpaughNLP & LLMs
27doc-assistant-hfML Application · Jupyter Notebook53.7/100ImplementationPop AlexandruNLP & LLMs
28promptlintAI Developer Tool · Pythonpromptlint is a static analysis linter for LLM prompts that detects common anti-patterns such as prompt injection risks, missing token limits, and hardcoded secrets.53.6/100ImplementationPop AlexandruNLP & LLMs
29llm-trainerML Application · JavaScriptllm-trainer is a full-stack application that provides a graphical interface for browsing Hugging Face datasets and fine-tuning local LLMs. It consists of a React frontend, an Express backend with WebSocket support, and a Python training worker that uses the Hugging Face…53.0/100ImplementationMaster MindNLP & LLMs

What it is

Where it counts

Around it on the NLP & LLMs board

More from Pop Alexandru