|
{ |
|
"base_current_gpu_type": "NVIDIA A100-PCIE-40GB", |
|
"base_current_gpu_total_memory": 40339.3125, |
|
"base_perplexity": 7.192307949066162, |
|
"base_token_generation_latency_sync": 35.781991958618164, |
|
"base_token_generation_latency_async": 36.746449396014214, |
|
"base_token_generation_throughput_sync": 0.027947018744973698, |
|
"base_token_generation_throughput_async": 0.027213513589382794, |
|
"base_token_generation_CO2_emissions": null, |
|
"base_token_generation_energy_consumption": null, |
|
"base_inference_latency_sync": 123.88730926513672, |
|
"base_inference_latency_async": 39.12379741668701, |
|
"base_inference_throughput_sync": 0.008071851797667635, |
|
"base_inference_throughput_async": 0.025559891064497787, |
|
"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": 8.009126663208008, |
|
"smashed_token_generation_latency_sync": 58.28822822570801, |
|
"smashed_token_generation_latency_async": 57.91718382388353, |
|
"smashed_token_generation_throughput_sync": 0.017156122778817805, |
|
"smashed_token_generation_throughput_async": 0.017266032876198413, |
|
"smashed_token_generation_CO2_emissions": null, |
|
"smashed_token_generation_energy_consumption": null, |
|
"smashed_inference_latency_sync": 200.24903717041016, |
|
"smashed_inference_latency_async": 110.69650650024414, |
|
"smashed_inference_throughput_sync": 0.004993781813537554, |
|
"smashed_inference_throughput_async": 0.009033708755730195, |
|
"smashed_inference_CO2_emissions": null, |
|
"smashed_inference_energy_consumption": null |
|
} |