# Model Training

> Free resources, with credit to the original creators.

## Resources
- [LlamaFactory](/model-training/fine-tuning/hiyouga-llamafactory/index.md): Compare adaptation methods within a shared training interface. (advanced; CN; open resource; by hiyouga (Yaowei Zheng) et al.)
- [TRL - Transformers Reinforcement Learning](/model-training/post-training-rl/hugging-face-trl-transformers-reinforcement-learning/index.md): Find training interfaces for supervised and preference-based adaptation. (advanced; other; closed resource; by Hugging Face)
- [Reinforcement Learning from Human Feedback (book)](/model-training/post-training-rl/nathan-lambert-reinforcement-learning-from-human-feedback/index.md): A conceptual path through reward modeling and post-training methods. (advanced; West; closed resource; by Nathan Lambert)
- [Controlling Reasoning Effort in LLMs (Ahead of AI)](/model-training/post-training-rl/sebastian-raschka-ahead-of-ai/index.md): A free Ahead of AI article on training controllable reasoning effort. (advanced; West; closed resource; by Sebastian Raschka)
- [Unsloth](/model-training/fine-tuning/unsloth-unsloth/index.md): Study local training and adaptation workflows for supported models. (advanced; West; open resource; by Unsloth)
- [nanochat](/model-training/pretraining/andrej-karpathy-nanochat/index.md): A compact training harness spanning the model-development lifecycle. (expert; West; open resource; by Andrej Karpathy)
- [DeepSeek-R1: Incentivizing Reasoning via RL (arXiv 2501.12948)](/model-training/post-training-rl/deepseek-deepseek-r1-incentivizing-reasoning-via-rl/index.md): Study the reported reinforcement-learning recipe behind DeepSeek-R1. (expert; CN; closed resource; by DeepSeek)
- [The Smol Training Playbook](/model-training/pretraining/hugging-face-the-smol-training-playbook/index.md): A training account that includes failed experiments and engineering decisions. (expert; other; closed resource; by Hugging Face (SmolLM team))
- [The Ultra-Scale Playbook](/model-training/training-infrastructure/hugging-face-the-ultra-scale-playbook/index.md): Understand how parallelism changes memory use and training throughput. (expert; other; closed resource; by Hugging Face (nanotron))
- [The FineWeb Datasets](/model-training/training-data/huggingface-fineweb/index.md): Study how filtering and deduplication affect pretraining data. (expert; other; open resource; by Guilherme Penedo and coauthors, Hugging Face)
- [verl (HybridFlow)](/model-training/post-training-rl/verl-project-verl/index.md): Explore how an RL training framework coordinates generation and optimization. (expert; CN; open resource; by verl project (originated at ByteDance Seed))