{ "current_gpu_type": "NVIDIA A100-PCIE-40GB", "current_gpu_total_memory": 40339.3125, "perplexity": 10.666025161743164, "token_generation_latency_sync": 153.23092956542968, "token_generation_latency_async": 153.4851636737585, "token_generation_throughput_sync": 0.006526097589018407, "token_generation_throughput_async": 0.006515287706410224, "token_generation_CO2_emissions": null, "token_generation_energy_consumption": null, "inference_latency_sync": 249.45889282226562, "inference_latency_async": 182.7211618423462, "inference_throughput_sync": 0.004008676494497551, "inference_throughput_async": 0.0054728198415398156, "inference_CO2_emissions": null, "inference_energy_consumption": null }