List your company

Time Series Analysis on General Efficiency Profiling

0.104Training Time (s/iter)GPT-2(6)-768
0.03420.505350.97651.44765Feb 5, 2025

Evaluation results

MethodLinks
GPT-2(6)-768Parameter=82.28M, Batch size=1282025.020.1040.0545,2302,566
GTMParameter=35.73M, Batch size=1282025.020.290.1658,3241,250
FEDformer-768Parameter=30.75M, Batch size=1282025.020.4670.1729,5351,880.19
Time-Moe(base)Parameter=50.00M, Batch size=1282025.020.840.0951,812.48226.7
TimesNet-768Parameter=42.21M, Batch size=1282025.021.8490.54735,8711,904.18