|
{ |
|
"base_current_gpu_type": "NVIDIA A100-PCIE-40GB", |
|
"base_current_gpu_total_memory": 40339.3125, |
|
"base_perplexity": 7.6169915199279785, |
|
"base_token_generation_latency_sync": 33.30435447692871, |
|
"base_token_generation_latency_async": 32.676383294165134, |
|
"base_token_generation_throughput_sync": 0.030026103664394423, |
|
"base_token_generation_throughput_async": 0.03060314206127473, |
|
"base_token_generation_CO2_emissions": null, |
|
"base_token_generation_energy_consumption": null, |
|
"base_inference_latency_sync": 121.60829544067383, |
|
"base_inference_latency_async": 39.43202495574951, |
|
"base_inference_throughput_sync": 0.008223123236587478, |
|
"base_inference_throughput_async": 0.02536009756339414, |
|
"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": 8.250771522521973, |
|
"smashed_token_generation_latency_sync": 166.4607131958008, |
|
"smashed_token_generation_latency_async": 166.59034881740808, |
|
"smashed_token_generation_throughput_sync": 0.006007423498322645, |
|
"smashed_token_generation_throughput_async": 0.006002748701223103, |
|
"smashed_token_generation_CO2_emissions": null, |
|
"smashed_token_generation_energy_consumption": null, |
|
"smashed_inference_latency_sync": 267.20727081298827, |
|
"smashed_inference_latency_async": 197.58310317993164, |
|
"smashed_inference_throughput_sync": 0.00374241313478283, |
|
"smashed_inference_throughput_async": 0.005061161525990089, |
|
"smashed_inference_CO2_emissions": null, |
|
"smashed_inference_energy_consumption": null |
|
} |