Model customization recipes are live on our website and in the cookbook repo.
Beam support added across RL and SFT training stacks
Developer tool
Reflection AI has published model customization recipes on its website and in its cookbook repository, alongside support for Beam across reinforcement learning and supervised fine-tuning stacks.
- Beam now works with Axolotl, NeMo RL, NeMo AutoModel, Megatron Bridge, Miles, SkyRL and TRL, according to the company.
- The recipes and integrations are available now, with the company pointing readers to its cookbook repository to get started.
