Build & Run AI/ML Models on AMD MI300X Accelerators

Develop, train, and deploy AI applications on the cloud.

• 2.4x more memory capacity
• 1.6x more memory bandwidth
• 1.3x more FP16 TFLOPS
• Store larger models, such as Llama3 70B on a single GPU; (no parallelism required)
  
• Full support for PyTorch and Tensorflow