search

lj107024 / FDDWNet

FDDWNET: A LIGHTWEIGHT CONVOLUTIONAL NEURAL NETWORK FOR REAL-TIME SEMANTIC SEGMENTATION(ICASSP2020)
9 stars 3 forks source link

Trained weights #2

Open ofirkris opened 4 years ago

ofirkris commented 4 years ago

Thank you for your hard work, can you share trained weights?

lj107024 commented 4 years ago

The weight setting refers to other code, and weights are calculated through all the dataset images, which can be seen in the loss.py. weight = torch.FloatTensor( [2.8149, 6.9850, 3.7890, 9.9428, 9.7702, 9.5110, 10.3113, 10.0264, 4.6323, 9.5608, 7.8698, 9.5168, 10.3737, 6.6616, 10.2604, 10.2878, 10.2898, 10.4053, 10.13809])

ofirkris commented 4 years ago

Thanks, can you share the pre-trained model?