Training CodeParrot 🦜 from Scratch

KV Cache from scratch in nanoVLM

How to train a new language model from scratch using Transformers and Tokenizers

LoRA training scripts of the world, unite!

Training Design for Text-to-Image Models: Lessons from Ablations

Smol2Operator: Post-Training GUI Agents for Computer Use

Guiding LLM Post-training Data Engineering with Model Internals from Sparse Autoencoders

Model internals encode rich information about how a large language model (LLM) processes its training data; however, post-training data engineering largely relies on external signals and ignores rich intrinsic signals lying in model internals. We propose SAERL, a data engineering framework for LLM reinforcement lear...