EN
When we launched SambaNova Cloud two months ago, we set out to make it easy for developers to create their own generative AI applications using the largest and most capable models, including Llama 3.1 405B and the lightning-fast Llama 3.1 70B. In fact, when we launched SambaNova Cloud, we were an order of magnitude faster than cloud providers using GPUs. Since then, we have only gotten faster and added more models, such as Llama 3.2 3B and Llama 3.2 11B and 90B-vision-instruct. And SambaNova is still the only AI accelerator system that offers the Llama 3.1 405B model to developers for free.
Now, for the last few weeks, our friends at Gradio, the service that enables the easy creation and sharing of AI and Machine Learning apps via a Web interface, have taken that easy access a step further.
Working together, we’ve created a SambaNova-Gradio integration that can be used to build and deploy AI apps using the SambaNova API with only a few lines of code.
Gradio developers can now utilize SambaNova Cloud’s open-source models with fast inference in their chatbot applications. And to make it even easier, we’ve implemented two key features.
First, Gradio created a sambanova_gradio library that contains a lot of pre-built boilerplate. This reduces the amount of code developers need to produce to build their applications.
Second, just last week, we added a one-click-deploy to Hugging Face. This allows developers to have a working Gradio Chatbot live on Hugging Face in seconds, where end-users can interact with these open-source models hosted on SambaNova.
What this means to developers is that they can get their AI-powered applications up and running on the lightning fast SambaNova platform in less than a minute, as opposed to having to write code and deploy their app with an API provider. This reduces the time to deploy to under a minute, compared to an hour or more for the traditional approach which includes the need to have extensive API knowledge, deployment protocols, and having to read product documentation. Now it is as simple as clicking the “Deploy to Hugging Face” button.
It doesn’t get any easier than that.
You can read more about it at VentureBeat.
