search

lhoyer / DAFormer

[CVPR22] Official Implementation of DAFormer: Improving Network Architectures and Training Strategies for Domain-Adaptive Semantic Segmentation
Other
470 stars 93 forks source link

How to obtain parameters in your code. #86

Open yuheyuan opened 9 months ago

yuheyuan commented 9 months ago

Hi, I am particularly intrigued by the table you provided regarding the model parameters. I'm interested in understanding how these parameters were obtained. Could you please share the method or tool you used to calculate the model parameters? This would greatly assist me in my learning and comprehension.

Thank you for your time and consideration. I look forward to your response.

[ image

I try those methods to get the parameters in your code. But I can't get the 3.7M params in your table7. 

flops, params = profile(model, inputs=(input, ))
print(flops/1e9, params/1e6)