[DOC]: What is the hardware used in th Energon-AI inference acceleration comparison ?

[lz02k]

📚 The doc issue

In the Inference (Energon-AI) Demo, what is the hardware used in th Energon-AI inference acceleration ? Can you show the gpu memory?

[JThh]

Quoting answers from Slack: A100 and you may take a look at bloom (176b) model for memory usage (complete GPT3 training requires more memory than what our test environment can afford).