Google Deepmind's Gemma
Gemma is a family of lightweight, state-of-the art open models built from the research and technology used to create Gemini models. Gemma models demonstrate strong performance across academic benchmarks for language understanding, reasoning, and safety. There are two sizes of models (2 billion and 7 billion parameters), and both pretrained and fine-tuned checkpoints are released. Gemma outperforms similarly sized open models on 11 out of 18 text-based tasks.
In this video, I will talk about the following: What is the architecture of Gemma models? How do Gemma models perform?
For more details, please look at https://blog.google/technology/develo... and https://storage.googleapis.com/deepmi...
Gemma: Open Models Based on Gemini Research and Technology. Gemma Team, Google DeepMind. 2024.