Forge is a training framework that uses YAML configs to manage different model training experiments.
- Github Link: Forge
Detail
- Forge is a YAML-driven training framework, so experiments can be modified by changing configuration files rather than rewriting training code.
- It is built on the Hugging Face stack, using Transformers, Hugging Face Trainer, and TRL for training workflows.
- Transformers = load models, tokenizers, and processors
- Hugging Face Trainer = run the standard supervised training paths, including SFT and KD
- TRL = run the current RL-style training path, currently GRPO
- supports multiple training paradigms:
- supervised fine-tuning
- knowledge distillation
- reinforcement-learning-based post-training, currently via the GRPO path
- provides experiment infrastructure:
- W&B logging
- evaluation
- checkpointing
- resume-from-checkpoint
- SLURM auto-resubmission for long-running jobs