Keras introduces Gemma models, which are lightweight and state-of-the-art open models that come with new features specifically designed for large language models. Gemma models deliver significant improvements in various benchmarks. These models can be fine-tuned with LoRA and can be run on multiple GPUs/TPUs.
•4m read time• From developers.googleblog.com
Table of contents
Get startedFine-tuning Gemma Models with LoRAFine-tuning Gemma models on multiple GPU/TPUsWhat’s NextSort: