Google has launched Gemma, a new series of optimized AI models, in collaboration with NVIDIA. The partnership aims to leverage NVIDIA's TensorRT-LLM for enhanced performance on RTX GPUs, including the upcoming feature, Chat with RTX. Gemma is designed to work seamlessly with GPUs, and it's already supported by popular no-code fine-tuning tool 🤗 AutoTrain. The launch includes integration with Hugging Chat, a fine-tune script, availability of transformers and GGUF, a free Google Colab example, Flash Attention 2, and JAX weights. This development marks Google's entry into the open-source Large Language Model (LLM) space, offering tools and integrations for the community to build upon.