LLM AI IP
DeepTransformCore
transformer-based AI accelerator
For sophisticated workloads, DeepTransformCore is optimized for language and vision applications. Supporting both encoder and decoder transformer architectures with flexible DRAM configurations and FPGA compatibility, DeepTransformCore eliminates complex software integration burdens, empowering customers to rapidly develop custom AI SoC (System-on-chip) designs with unprecedented efficiency.
Supports DeepMentor Exclusive Low-distortion Training Models
