postgresml vs LLMs-from-scratch
A neutral, constraint-first comparison - live GitHub stats and typed relationships, not marketing.
| postgresml | LLMs-from-scratch | |
|---|---|---|
| Tagline | Postgres with GPUs for ML/AI applications. | Implement a ChatGPT-like LLM in PyTorch from scratch |
| Stars | 6.8k | 99k |
| Forks | 365 | 15k |
| Open issues | 155 | 4 |
| Language | Rust | Jupyter Notebook |
| License | MIT | Other |
| Last pushed | Jul 1, 2025 | Jun 2, 2026 |
| Categories | Inference & Serving, Model Training, LLM Frameworks | Model Training, LLM Frameworks |
postgresml
PostgresML is a powerful PostgreSQL extension that allows machine learning operations to be run directly within the database, combining data storage and inference. It supports GPU acceleration, integration of large language models from Hugging Face, and includes RAG pipeline functions for text processing.
Rust
LLMs-from-scratch
Repository containing code and instructions for developing, pretraining, and finetuning a GPT-like large language model (LLM) using PyTorch.
Jupyter Notebook