Load test different models in the inference-server

[jackapbutler]

We want to test the performance of different models within the inference server to understand how it scales with model size such as;

• distilgpt2
• pythia-12B