Want a smart insight into your inbox? Sign up for our weekly newsletters to get the only thing that is important to enterprise AI, data, and security leaders. Subscribe now
Liquid AI, the former Massachusetts Institute of Technology (MIT), is created by researchers who produce novel AI model architecture beyond widely used.Transformer“, Today Announced the release of Jump The alias “Liquid Edge AI Platform,” a cross -platform software development kit (SDK) has been developed to make developers easier to integrate smaller language models (SLMS) directly into mobile applications.
Along with the leap, the company also introduced ApolloLocally, a fellow iOS app to test these models, which can carry forward liquid AI’s mission to enable users’ hardware to protect, efficient AI.
Lip SD arrives at a time when many developers are looking for alternatives to AI services only due to delays, cost, privacy and offline availability concerns.
Leap them needs head with local first view Allows small models to run directly on deviceReduce dependence on cloud infrastructure.
Made for mobile Davis that does not require a pre -ML experience
Leaps are designed for developers who want to build with AI, but it does not have a deep skill in machine learning.
According to Liquid AI, SDK can be added to the iOS or Android project that has only a few lines of code, and the purpose of calling the local model is to be familiar with the conversation with the traditional cloud API.
“Our research shows that developers clouds are just moving beyond AI and are looking for reliable partners to help them make a device.” The website of the liquid. “LEAP is our answer. It is a flexible, from the end deployment platform that is made from the ground to make the powerful, efficient, and private edge really accessible.”
Once connected, developers can choose a model from the Built-in Leap Model Library, which includes more than 300MB of compact models-the lightweight for the innocent phone is so low that 4GB RAM (!!) and more.
SD’s local estimates, memory correction, and device compatibility, facilitates the process of deployment of the normal edge.
OS-and Model-Egonostic in terms of lip design. At the time of launch, it supports both iOS and Android, and compattate the liquid AI’s own liquid foundation model (LFMs) as well as many famous open source small models.
Purpose: A united environmental system for Edi AI
Beyond the implementation of the model, the leap is positioned as an All -In One Platform to discover, mold, test, and deploy the SLS SLS SLS SLS SLS.
Developers can browse a curatic model catalog with various quantization and outpost options, which can make them prepare performance and memory image according to the target tool obstacles.
Liquid AI emphasizes that large models are usually, while small models often perform a tight set of tasks to optim better. The unified system of lip is formed around this principle, which offers tools tools for rapid repetition and deployment in the real world’s mobile environment.
The SD’s Discover also comes with a developer community, where the liquid AI office offers time, help, events and competitions to encourage experiences and feedback.
Apollo: Like Test flight For local AI models
For the completion of the jump, the liquid AI was also released ApolloA free iOS app that allows developers and users to interact with the model in a local, offline sequence.
OriginalMobile App Startup Apps that allow users to chat with LLM privately on the deviceWhich liquid acquired earlier this year, Apollo has been rebuilt to support the entire leap model library.
Apollo is designed for low friction experiments-the driver can “check” a model tone, delay, or output behavior on his phone before integrating into a production app. The app operates completely offline, which protects user privacy and reduces dependence on cloud computations.
Whether used as a lightweight giant tool or a private AI assistant, reflects the wider pressure of liquid AI to access Apollo AI and to execute the process.
The LFM 2 Model Family announced last week is built on the back of the Family
Lip SD’s release continues Liquid AI announces July 10 Lfm2The Foundation’s Foundation’s Family Family is specifically designed for on -device workload.
LFM 2 models come in 350 meters, 700 meters, and 1.2B parameters, and benchmarks match with large models in speed and accuracy in multiple diagnostic works.
These models create the LEAP Model Library’s backbone and are better at CPU, GPU, and NPU.
Are free and ready to start construction for Davis
Leap is currently free to use under a developer license, which includes the Core SD and the Model Library.
Liquid AI notes that the features of the premium enterprise in the future will be made available under a separate trading license, but it is already questioning the enterprise users. Through its website contact form.
LFM2 models are also free for educational use and commercial use by less than $ 10 million in income companies, large organizations need to contact the company for licensing.
Developers can start by visiting Leap SD’s websiteDownloading Apollo from the App Store, or joining the liquid AI developer community on the discard.
