IBM recently announced that its AI development platform watsonx.ai now supports the distilled versions of the Llama3.18B and Llama3.370B models, known as DeepSeek-R1. DeepSeek optimizes various Llama and Qwen variants using data generated by the R1 model through knowledge distillation technology, further enhancing model performance.
On the watsonx.ai platform, users can utilize the DeepSeek distilled models in two ways. First, IBM offers the Llama distilled version in the "On-Demand Deployment" catalog, allowing users to deploy dedicated instances to ensure secure inference. Secondly, users can also import other variants of DeepSeek-R1, such as the Qwen distilled model, through the "Custom Base Model" upload feature to meet diverse application needs.
DeepSeek-R1 possesses powerful inference capabilities, making it suitable for a wide range of fields and providing efficient and flexible AI solutions for enterprises and developers. This update further enriches the model ecosystem of watsonx.ai, enabling users to develop and deploy AI applications more conveniently.
