search

tsing90 / pytorch_semantic_human_matting

This is an unofficial implementation of the paper "Semantic human matting":
https://arxiv.org/pdf/1809.01354.pdf
83 stars 18 forks source link

Train time and Train loss #10

Closed kkkmax closed 5 years ago

kkkmax commented 5 years ago

@tsing90 thanks! I prepared the data according to the description of your Readme file, re-prepared the data, 180 DIM images and my own 880 collected images were processed into four-channel (RGBA) images, currently training, ask a few Questions about model training: 1) How long have you trained T-Net training? How much do you think the loss will fall to get a robust T-Net model? 2) How long did it take to train M-net, and how long did you train in end-to-end mode?

tsing90 commented 5 years ago
  1. my datasets are larger than yours, it took 50-100 epochs to converge, if you add more weights on unknown area, it can be faster;
  2. m-net is similar; but my end-to-end was not stable, and I haven't figured out the reason, so currently I haven't finished yet.