how many epoch do you train when finetuning on IC15 about craft-reimplementation HOT 26 OPEN

thanks， i fixed it. i put net.train() at start and net.eval() at end of 'for' code block. it seems to work.

from craft-reimplementation.

what's the hmean value on IC15 using pretrained model on synthtext.

from craft-reimplementation.

@kouxichao IC15 for 500 epochs,MLT for ~65 epochs. I forgot the IC15 hmean, maybe about 56-57%.

from craft-reimplementation.

why is the hmean value of IC15 still very low , and go up and down between 0-0.3, after traing on IC15 for 200 epoch? is that normal?

from craft-reimplementation.

@kouxichao Abnormal, it should be about ~80% for 200 epoch.

from craft-reimplementation.

你使用IC15进行训练的时候，，，hmean怎么变化的？ 刚开始是0.0几，0.1几的情况吗? 我训练的时候为什么hmean一直在震荡，有时候0.1，有时候0.2。

from craft-reimplementation.

@kouxichao could you share you hmean?

from craft-reimplementation.

this is the loss figure generated in training process：

the hmean value i didn't save. screenshots of hmean on IC15:

i retrained on IC15 for several times on IC15 for 100-200epoch， but the hmeans always are like above, i changed the batchsize to 1 for synthtext, and 5 for IC15, because the memory is not enough.
(i test the batchsize 2 for synthtext, and 5 for IC15 for 100 epoch , hmean is still like that. for memory reason , i only can train 100 epoch using original batchsize.)

from craft-reimplementation.

@kouxichao I will check it， and tomorrow give you the reason.

from craft-reimplementation.

the output map and result in traing process.

thanks, hope to see your check results!

from craft-reimplementation.

@kouxichao I am so sorry for that I am so busy today, I will check it at noon tomorrow.

from craft-reimplementation.

never mind, can you share your train info(the hmean and loss of every epoch)?, so that i can find something wrong happened much earlier in training.
or can you just simply describe the trend of hmean and loss change. the hmean are always greater than the hmean(0.58) evaluted on pretrained model， or greater than some other value like 0.4?

And this pic is the input of net, is that normal?

from craft-reimplementation.

@kouxichao Have you set the model to eval mode when generating the pseudo label code?

from craft-reimplementation.

@lianqing11 you mean this eval mode?

from craft-reimplementation.

@lianqing11 thanks for @kouxichao ,just use the line 528 code, it will work.

from craft-reimplementation.

you mean uncomment this line? i retrain with this uncommented, the hmean of epoch 0-17 is still like before, is that normal?

from craft-reimplementation.

@kouxichao data_loader.py file

from craft-reimplementation.

@kouxichao Have you also set the model to train model --> model.train() after generating pseudo bounding box in every iteration, which may solve the problem.

from craft-reimplementation.

you mean put net.train() in for loop?

from craft-reimplementation.

@kouxichao I am checking it, just for a moment.

from craft-reimplementation.

@kouxichao sorry， there is no leisure GPU， so i can not check it. Or can you train MLT, first?

from craft-reimplementation.

@kouxichao ok，sorry for the problems.

from craft-reimplementation.

thanks for your great work and patience with the problem.

from craft-reimplementation.

for training epoch, the author trained the model for 16 images one GPU, so for weakly supervision, if you train the model for real image and syndata image with 10:2, for IC15 you should train 500 epoch, for MLT you should keep the iteartions * epoch = 25K.

from craft-reimplementation.

net.train()
could you please share this part of code? I mean net.train() and net.eval(). is it possible to use MLT pretrained weight for training custom dataset?

from craft-reimplementation.

Hi @kouxichao, I also having similar issue but even worse result on 143th epoch:

Loading weights from checkpoint ./data/CRAFT-pytorch/real_weights/CRAFT_clr_143.pth
elapsed time : 86.13623404502869spytorch/test/img_85.jpgg
Calculated!{"precision": 0.0008183306055646482, "recall": 0.00048146364949446316, "hmean": 0.0006062443164595332, "AP": 0}

Could you share that which part of the code you have modified to fix the problem?
Thanks

from craft-reimplementation.