Fine-tuning SqueezeNet - Githubissues

forresti / SqueezeNet

SqueezeNet: AlexNet-level accuracy with 50x fewer parameters

BSD 2-Clause "Simplified" License

2.17k stars 723 forks source link

Closed mun3 closed 6 years ago

mun3 commented 6 years ago

When fine-tuning SqueezeNet, should some layers be frozen?

Thanks

parcar commented 5 years ago

Did you find the answer to your question? I would be interested in the answer if so.

syedmustafan commented 3 years ago

@parcar Remove the RELU layer and nothing more.