inference-cache-config/gpt2.json

{
    "openai-community/gpt2": [
    {
      "batch_size": 1,
      "sequence_length": 1024,
      "num_cores": 1,
      "auto_cast_type": "fp16"
    },
    {
      "batch_size": 16,
      "sequence_length": 1024,
      "num_cores": 1,
      "auto_cast_type": "fp16"
    }
  ]
}