test / gpu_training /eval /.ipynb_checkpoints /summary-checkpoint.txt
tanglumy
initial commit of model weights
1a6c70c
FR-Ponder Evaluation Results
==============================
Dataset: gsm8k
--------------------
FR-Ponder:
Accuracy: 0.000
Avg FLOPs: 1548841728
Avg Time: 3.288s
Avg Steps: 118.1
Baseline (α=0.4):
Accuracy: 0.240
Avg FLOPs: 1264988628340
Avg Time: 5.507s
Improvements:
Accuracy: +-0.240
FLOPs reduction: 99.9%
Speedup: 1.68x