Table 2 User-defined and evaluation based parameters.
From: An automated multi parameter neural architecture discovery framework using ChatGPT in the backend
Legends | Description | Type |
|---|---|---|
\(T_{Acc}, V_{Acc}\) | Current model’s training and validation accuracy respectively as predicted by LLM/AGI | Evaluation based |
\(PT_{Acc}, PV_{Acc}\) | Priority of training and validation accuracy respectively | User defined |
\(TT_{Acc}, TV_{Acc}\) | Threshold of the training and validation accuracy | User defined |
\(T_E, V_E\) | Energy required for evaluating the training and validation set | Evaluation based |
\(PT_E, PV_E\) | Priority of the energy required for evaluating the training and validation set | User defined |
\(TT_E, TV_E\) | Threshold of energy required for evaluating the training and validation set | User defined |
F | FPS of the current model predicted by LLM/AGI | Evaluation based |
NF | Normalized FPS of the current model predicted by LLM/AGI | Evaluation based |
PF | Priority of the FPS for the model | User defined |
TF | Threshold ofthe FPS for the model | User defined |
P | Parameters of the current model (CM) predicted by LLM/AGI | Evaluation based |
OT, UT | Threshold value to check the overfitting and underfitting | User defined |
\(W_A\) | Weight for the accuracy values when computing the combined metric (CM) | User defined |
\(W_E\) | Weight for the energy values when computing the combined metric (CM) | User defined |
\(W_F\) | Weight for the FPS values when computing the combined metric (CM) | User defined |
