MusicFX Integrations

MusicLM is an experimental AI tool which can convert your text descriptions into musical compositions. MusicLM will create the two versions of a song based on a prompt such as "soulful Jazz for a Dinner Party". You can listen to them both and award a trophy to whichever track you prefer. This will help improve the model. We believe responsible innovation doesn't happen in isolation. We've worked with musicians such as Dan Deacon, and hosted workshops to see how technology can empower creative processes. MusicLM can help you express creativity, whether you're a professional or just starting out.

We're beta-launching SynthID, an AI-generated image watermarking tool. SynthID will be released to a small number of Vertex customers who use Imagen, our latest text-to image model that uses input text in order to create photorealistic pictures. This tool allows users to embed a digital watermark that is imperceptible into their AI-generated image and identify whether Imagen was used to generate the image or if a part of it. To promote trust in information, it is important to be able to identify AI generated content. SynthID, while not a panacea for the problem of misinformation and false information, is a promising early solution to this pressing AI issue. This technology was developed and refined by Google DeepMind in partnership with Google Research. SynthID can be used with other AI models, and we plan to incorporate it into more products soon.

Chinchilla has a large language. Chinchilla has the same compute budget of Gopher, but 70B more parameters and 4x as much data. Chinchilla consistently and significantly outperforms Gopher 280B, GPT-3 175B, Jurassic-1 178B, and Megatron-Turing (530B) in a wide range of downstream evaluation tasks. Chinchilla also uses less compute to perform fine-tuning, inference and other tasks. This makes it easier for downstream users to use. Chinchilla reaches a high-level average accuracy of 67.5% for the MMLU benchmark. This is a greater than 7% improvement compared to Gopher.