|
{ |
|
"base_current_gpu_type": "NVIDIA A100-PCIE-40GB", |
|
"base_current_gpu_total_memory": 40339.3125, |
|
"base_perplexity": 8.711685180664062, |
|
"base_token_generation_latency_sync": 38.0333869934082, |
|
"base_token_generation_latency_async": 38.08097634464502, |
|
"base_token_generation_throughput_sync": 0.026292688583672973, |
|
"base_token_generation_throughput_async": 0.02625983091792815, |
|
"base_token_generation_CO2_emissions": null, |
|
"base_token_generation_energy_consumption": null, |
|
"base_inference_latency_sync": 120.86487045288087, |
|
"base_inference_latency_async": 39.441466331481934, |
|
"base_inference_throughput_sync": 0.008273702658621967, |
|
"base_inference_throughput_async": 0.025354026941990394, |
|
"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": 9.663578033447266, |
|
"smashed_token_generation_latency_sync": 62.078622055053714, |
|
"smashed_token_generation_latency_async": 61.78989615291357, |
|
"smashed_token_generation_throughput_sync": 0.016108604973756692, |
|
"smashed_token_generation_throughput_async": 0.016183875718536016, |
|
"smashed_token_generation_CO2_emissions": null, |
|
"smashed_token_generation_energy_consumption": null, |
|
"smashed_inference_latency_sync": 196.9707015991211, |
|
"smashed_inference_latency_async": 111.20069026947021, |
|
"smashed_inference_throughput_sync": 0.005076897182583129, |
|
"smashed_inference_throughput_async": 0.008992749933266797, |
|
"smashed_inference_CO2_emissions": null, |
|
"smashed_inference_energy_consumption": null |
|
} |