search

QiaoranC / tf_ResNeSt_RegNet_model

tensorflow 2.x version of ResNeSt,RegNet,DETR
129 stars 31 forks source link

Any hope of having pretrained weights ? #5

Open 0xsimulacra opened 4 years ago

0xsimulacra commented 4 years ago

Hello,

Thanks for sharing your ready to train code, I was wondering if there any hope to have the pretrained weights on Image-Net ? I can contribute by adding Cascade-RCNN and Faster-RCNN to the model as having these pretrained weights will help me to test the performance of the model on a task without the need to learn Pytorch as I don't have eanough time.

If there no chance, how long do you think it will take me to train the model (Resnest-200) on Colab pro ?

Thank you.

QiaoranC commented 4 years ago

Hi, the pertrain model is indeed a reasonable ask, but i dont have much time/power/gpu to do it, Sorry for that. And welcome to contribute it. As for the time of Colab pro, sorry i never use them, cannt give you any suggestions. I leave this open. And again, welcome anyone contribute the pertrain model or the usage feedback or improvements/suggests.

0xsimulacra commented 4 years ago

Thank you for the clarification, I assumed you just didn't want to release the weights. I too have only the limited power of Colab Pro (1 x 16GB GPU) and the other problem is that I'm not that experienced to pretrain a model on Benchmark dataset. If I find I can do it (in the required power side and time/reading needed to program the learning rate scheduleer, the data processing and the training strategy) I'll let you know.