|
{ |
|
"base_current_gpu_type": "NVIDIA A100-PCIE-40GB", |
|
"base_current_gpu_total_memory": 40339.3125, |
|
"base_perplexity": 8.629734992980957, |
|
"base_token_generation_latency_sync": 35.20570373535156, |
|
"base_token_generation_latency_async": 35.54570898413658, |
|
"base_token_generation_throughput_sync": 0.028404488304429404, |
|
"base_token_generation_throughput_async": 0.028132790949430275, |
|
"base_token_generation_CO2_emissions": null, |
|
"base_token_generation_energy_consumption": null, |
|
"base_inference_latency_sync": 119.19165420532227, |
|
"base_inference_latency_async": 38.88437747955322, |
|
"base_inference_throughput_sync": 0.008389849160725441, |
|
"base_inference_throughput_async": 0.02571726911471928, |
|
"base_inference_CO2_emissions": null, |
|
"base_inference_energy_consumption": null, |
|
"smashed_current_gpu_type": "NVIDIA A100-PCIE-40GB", |
|
"smashed_current_gpu_total_memory": 40339.3125, |
|
"smashed_perplexity": 9.04076099395752, |
|
"smashed_token_generation_latency_sync": 43.718316650390626, |
|
"smashed_token_generation_latency_async": 42.23632402718067, |
|
"smashed_token_generation_throughput_sync": 0.02287370778698692, |
|
"smashed_token_generation_throughput_async": 0.02367630287513805, |
|
"smashed_token_generation_CO2_emissions": null, |
|
"smashed_token_generation_energy_consumption": null, |
|
"smashed_inference_latency_sync": 162.82562713623048, |
|
"smashed_inference_latency_async": 70.3972578048706, |
|
"smashed_inference_throughput_sync": 0.006141539373057874, |
|
"smashed_inference_throughput_async": 0.014205098766372866, |
|
"smashed_inference_CO2_emissions": null, |
|
"smashed_inference_energy_consumption": null |
|
} |