GuppyLM vs MinerU2.5

GuppyLM is a deliberately tiny language model — 9 million parameters, 6 transformer layers — that roleplays as a fish and can be fully trained in under 5 minutes on a free Google Colab T4 GPU. The entire pipeline from data generation to training loop to inference fits in approximately 130 lines of PyTorch, making it the most compressed end-to-end LLM tutorial available. Unlike educational projects that paper over complexity with abstraction layers, GuppyLM deliberately avoids modern optimizations — no RoPE positional encoding, no grouped-query attention, no SwiGLU activations. You see exactly why each component exists when you remove it. It ships with a 60,000-example synthetic conversation dataset and produces coherent (if goofy) fish-themed responses after training. The project hit the top of Hacker News Show HN with 365 points and 31 comments. Developers praised how the simplicity forces you to confront how training data shapes model behavior directly, with multiple commenters saying it's the clearest path from 'I know Python' to 'I understand why LLMs work.'

MinerU2.5 is a 1.2-billion parameter vision-language model purpose-built for high-resolution document parsing. From OpenDataLab, it's the latest version of a project that's accumulated 61.5K GitHub stars — which tells you something about how painful document-to-text has been as a category. The model uses a decoupled vision-language architecture for efficient high-resolution processing with state-of-the-art recognition accuracy across tables, formulas, figures, and mixed-layout documents. The core use case is turning messy PDFs, scanned forms, academic papers, and enterprise documents into clean Markdown or structured JSON that LLMs can actually work with. Earlier MinerU versions were already widely adopted for RAG pipeline preprocessing — 2.5 tightens up accuracy on the edge cases that killed earlier tools: rotated pages, dense tables, multi-column layouts, and multilingual content. At 1.2B parameters it's lightweight enough to run locally without a GPU farm, and the Apache 2.0 license means it integrates cleanly into commercial document pipelines. For anyone building RAG applications, AI research assistants, or document intelligence products, this is the preprocessing layer that removes a persistent pain point.