{ "current_gpu_type": "NVIDIA A100-PCIE-40GB", "current_gpu_total_memory": 40339.3125, "perplexity": 2305.849853515625, "token_generation_latency_sync": 155.22875213623047, "token_generation_latency_async": 155.3879875689745, "token_generation_throughput_sync": 0.006442105513560973, "token_generation_throughput_async": 0.006435503899914492, "token_generation_CO2_emissions": null, "token_generation_energy_consumption": null, "inference_latency_sync": 243.84829406738282, "inference_latency_async": 198.03030490875244, "inference_throughput_sync": 0.004100910378826227, "inference_throughput_async": 0.005049732163270544, "inference_CO2_emissions": null, "inference_energy_consumption": null }