search

GengDavid / pytorch-cpn

A PyTorch re-implementation of CPN (Cascaded Pyramid Network for Multi-Person Pose Estimation)
GNU General Public License v3.0
485 stars 99 forks source link

Training with other configurations. #3

Open mkocabas opened 6 years ago

mkocabas commented 6 years ago

Hi @GengDavid,

Thanks for the great implementation. I'm eager collaborate with you to test other configurations. I have 2 x 1080 and 2 x 1080ti. I can borrow more if needed. Looking forward to your response!

GengDavid commented 6 years ago

Hi @mkocabas ,

Thanks for your interest in my implementation. There may be at least two configurations to be tested, ResNet-50+384x288 and ResNet-101+384x288. Which one do you prefer to test? Or do you want to test both of them?

I've modified the codes a little, so please clone/pull the latest version before you run it. Please follow README to configure the environment.

You can train a ResNet-50+384x288 model directly in 384.288.model dir. by running train.py You may need to modify batch size in config.py, and use -g to specify the number of GPU you use. For example, you may set batch_size = 12 and run python3 train.py -g 2 when you use 2 x 1080 gpu to train the model.

To train a ResNet-101+384x288 model, you need to set model='CPN101' in config.py, and then follow the same way to train the model.

If you have any questions, feel free to contact me. You can also mail me at zhanggw8@mail2.sysu.edu.cn or zgwdavid@gmail.com.

mkocabas commented 6 years ago

Cool, so I can start with ResNet-50+384x288. After that I can try ResNet-101.

I'll use 2 x 1080ti with the default hyperparameters as in config. Am I correct?

mkocabas commented 6 years ago

@GengDavid we have a little problem. 1080tis have 11GB memory. batch_size=6 barely fits the memory. This means that we can train with batch_size=12 using 2 gpus. What do you think?

GengDavid commented 6 years ago

If you are using 1080tis, I think you can set batch_size more than 12 with 2 gpus while running ResNet-50+384x288 model.

GengDavid commented 6 years ago

@mkocabas ResNet-50+384x288 model with batch_size=12 takes about 8G memory in my experiment.

mkocabas commented 6 years ago

I'm consistently getting OOM error, but let me check. I'll restart the computer, maybe there are some blocking processes. I'll inform you about the progress.

mkocabas commented 6 years ago

@GengDavid, restarting solved the problem. Thanks for pointing out! I'll update this issue as training continues.

How many epochs did you train the 256x192 model?

GengDavid commented 6 years ago

@mkocabas About 25 epoch. I don't remember the exact figure.

mkocabas commented 6 years ago

I see, so probably it'll take 4 days to converge.

GengDavid commented 6 years ago

Fine, thanks.

mkocabas commented 6 years ago

Epoch 6 (tested with GT bboxes)

 Average Precision  (AP) @[ IoU=0.50:0.95 | area=   all | maxDets= 20 ] = 0.688
 Average Precision  (AP) @[ IoU=0.50      | area=   all | maxDets= 20 ] = 0.894
 Average Precision  (AP) @[ IoU=0.75      | area=   all | maxDets= 20 ] = 0.750
 Average Precision  (AP) @[ IoU=0.50:0.95 | area=medium | maxDets= 20 ] = 0.654
 Average Precision  (AP) @[ IoU=0.50:0.95 | area= large | maxDets= 20 ] = 0.742
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets= 20 ] = 0.719
 Average Recall     (AR) @[ IoU=0.50      | area=   all | maxDets= 20 ] = 0.904
 Average Recall     (AR) @[ IoU=0.75      | area=   all | maxDets= 20 ] = 0.776
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=medium | maxDets= 20 ] = 0.681
 Average Recall     (AR) @[ IoU=0.50:0.95 | area= large | maxDets= 20 ] = 0.777

Epoch 13 (tested with GT bboxes)

 Average Precision  (AP) @[ IoU=0.50:0.95 | area=   all | maxDets= 20 ] = 0.726
 Average Precision  (AP) @[ IoU=0.50      | area=   all | maxDets= 20 ] = 0.914
 Average Precision  (AP) @[ IoU=0.75      | area=   all | maxDets= 20 ] = 0.785
 Average Precision  (AP) @[ IoU=0.50:0.95 | area=medium | maxDets= 20 ] = 0.690
 Average Precision  (AP) @[ IoU=0.50:0.95 | area= large | maxDets= 20 ] = 0.781
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets= 20 ] = 0.754
 Average Recall     (AR) @[ IoU=0.50      | area=   all | maxDets= 20 ] = 0.924
 Average Recall     (AR) @[ IoU=0.75      | area=   all | maxDets= 20 ] = 0.810
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=medium | maxDets= 20 ] = 0.716
 Average Recall     (AR) @[ IoU=0.50:0.95 | area= large | maxDets= 20 ] = 0.812
mkocabas commented 6 years ago

@GengDavid do you have the weights of 5th epoch of ResNet50-256x192 model?

GengDavid commented 6 years ago

Yes, I do have saved the 5th epoch pre-trained model. But I'm sorry to tell you that there's something different from the original paper in my code just as @Tiamo666 mentioned in issue #4.
The results seem very close, but I'm still going to modify the network and then re-test it.

mkocabas commented 6 years ago

Yeah I saw the discussion. Please let me know about the results after modification. If you don't have enough GPUs, I can test the corrected model.

GengDavid commented 6 years ago

I'll let you know the results but it may take a little long time since I only have 1*1080 free to run the code. May be you can test test the ResNet-50+384x288 model first.
Thanks!

mkocabas commented 6 years ago

I've started to train fixed ResNet-50+384x288 on a Titan V w batch-size=24

GengDavid commented 6 years ago

Hi, @mkocabas I've updated the ResNet-50+256*192 results. Have got some results? Thx.

Tiamo666 commented 6 years ago

Hi, David, I've trained with the ResNet-50+384*288 with ground truth bboxes. The test result of 32 epoch is as follows: Average Precision (AP) @[ IoU=0.50:0.95 | area= all | maxDets= 20 ] = 0.737 Average Precision (AP) @[ IoU=0.50 | area= all | maxDets= 20 ] = 0.915 Average Precision (AP) @[ IoU=0.75 | area= all | maxDets= 20 ] = 0.806 Average Precision (AP) @[ IoU=0.50:0.95 | area=medium | maxDets= 20 ] = 0.706 Average Precision (AP) @[ IoU=0.50:0.95 | area= large | maxDets= 20 ] = 0.792 Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 20 ] = 0.767 Average Recall (AR) @[ IoU=0.50 | area= all | maxDets= 20 ] = 0.929 Average Recall (AR) @[ IoU=0.75 | area= all | maxDets= 20 ] = 0.826 Average Recall (AR) @[ IoU=0.50:0.95 | area=medium | maxDets= 20 ] = 0.729 Average Recall (AR) @[ IoU=0.50:0.95 | area= large | maxDets= 20 ] = 0.824

Due to the limit of network, I can not download the person detections results on COCO successfully, So I just use the ground truth.

GengDavid commented 6 years ago

@Tiamo666 Great job! Can you provide the pre-trained model so that I can test it with detection results? I think you can open a PR with the a link on it to download pre-trained model.

GengDavid commented 6 years ago

@Tiamo666 Or if you do not want to open a RP, could you just provide a link to download the model? Google Drive, Onedrive, Dropbox and Baidu Yun are all fine.

Tiamo666 commented 6 years ago

OK，I guess Baidu yun is a good choice. I will try to share the pretrained model on it and provide you the link as soon as I uploaded model

Tiamo666 commented 6 years ago

hi，David, I've already uploaded the model on BaiduYun. Here is the link: https://pan.baidu.com/s/1fdy5_0HQm63QtlOzxKbpuw

GengDavid commented 6 years ago

Great! I'll test it and update the result later.

GengDavid commented 6 years ago

@Tiamo666 I've updated the results.

Tiamo666 commented 6 years ago

That's cool! I'll have time to train with Resnet101+384*288, I'll share the model after finishing training

GengDavid commented 6 years ago

@Tiamo666 That's great! If you have any problem, feel free to contact me.

Tiamo666 commented 6 years ago

Hi, David. I've uploaded the model of cpn384*288 with Resnet101 on Baidu Yun. Here is the link: https://pan.baidu.com/s/1toikUHSqHhHP3DkIOkNctA

GengDavid commented 6 years ago

@Tiamo666 Great! Thanks a lot. I'll update the results soon.

Tiamo666 commented 6 years ago

Hello, David, I've just found that I trained with the old code which has "Color Normalized bug" last week. I feel sorry for that, I could retrain the model this week.

GengDavid commented 6 years ago

@Tiamo666 Retraining it is a better choice but may cost more time. I think we can just fine-tune the trained model. This may influence the result a little but can save time. However, I currently do not have free GPUs to do this work. What do you think about that?

Tiamo666 commented 6 years ago

OK, Thank you for your advice. I think fine-tune the model is a good idea. Another thing I wanted to mention is that in issue#7, it doesn't matter whether there is bias in nn.conv2d cause the batchnorm will minus the mean value, so plus a constant will not affect the result.

GengDavid commented 6 years ago

@Tiamo666 Yep, the bias has little influence to the result. However, it is better to avoid adding bias to conv2d with batchnorm.

GengDavid commented 6 years ago

@Tiamo666 Here is what I did. You can modify the training codes like this(from line 38)

    if args.resume:
        if isfile(args.resume):
            print("=> loading checkpoint '{}'".format(args.resume))
            checkpoint = torch.load(args.resume)
            checkpoint_state_dict = checkpoint['state_dict']
            new_dict = {}
            for k,v in checkpoint_state_dict.items():
                if k=='module.global_net.upsamples.0.1.bias':
                    continue
                if k=='module.global_net.upsamples.1.1.bias':
                    continue
                if k=='module.global_net.upsamples.2.1.bias':
                    continue
                new_dict[k]=v

            model.load_state_dict(new_dict)
            args.start_epoch = checkpoint['epoch']
            # optimizer.load_state_dict(checkpoint['optimizer'])

Using --resume to continue training, and set --epochs to one or two larger than the checkpoint you load. (and also make sure to change the learning rate to a proper value. )

Tiamo666 commented 6 years ago

Ok, That's cool, thanks a lot.

Tiamo666 commented 6 years ago

Hello, David. I've upload the model of cpn384x288 with resnet101 On BaiduYun. Here is the link: https://pan.baidu.com/s/1e_meK3xnGRZXJEBaFVXB3A

GengDavid commented 6 years ago

Cool. @Tiamo666 Could you please tell me the results you got before and after the fine-tune process(using gt bbox)?

Tiamo666 commented 6 years ago

Hello, David, the results after fine-tune is Average Precision (AP) @[ IoU=0.50:0.95 | area= all | maxDets= 20 ] = 0.740 Average Precision (AP) @[ IoU=0.50 | area= all | maxDets= 20 ] = 0.923 Average Precision (AP) @[ IoU=0.75 | area= all | maxDets= 20 ] = 0.806 Average Precision (AP) @[ IoU=0.50:0.95 | area=medium | maxDets= 20 ] = 0.711 Average Precision (AP) @[ IoU=0.50:0.95 | area= large | maxDets= 20 ] = 0.787 Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 20 ] = 0.770 Average Recall (AR) @[ IoU=0.50 | area= all | maxDets= 20 ] = 0.931 Average Recall (AR) @[ IoU=0.75 | area= all | maxDets= 20 ] = 0.829 Average Recall (AR) @[ IoU=0.50:0.95 | area=medium | maxDets= 20 ] = 0.736 Average Recall (AR) @[ IoU=0.50:0.95 | area= large | maxDets= 20 ] = 0.821 Before fine-tune is Average Precision (AP) @[ IoU=0.50:0.95 | area= all | maxDets= 20 ] = 0.075 Average Precision (AP) @[ IoU=0.50 | area= all | maxDets= 20 ] = 0.154 Average Precision (AP) @[ IoU=0.75 | area= all | maxDets= 20 ] = 0.063 Average Precision (AP) @[ IoU=0.50:0.95 | area=medium | maxDets= 20 ] = 0.100 Average Precision (AP) @[ IoU=0.50:0.95 | area= large | maxDets= 20 ] = 0.043 Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 20 ] = 0.084 Average Recall (AR) @[ IoU=0.50 | area= all | maxDets= 20 ] = 0.165 Average Recall (AR) @[ IoU=0.75 | area= all | maxDets= 20 ] = 0.073 Average Recall (AR) @[ IoU=0.50:0.95 | area=medium | maxDets= 20 ] = 0.109 Average Recall (AR) @[ IoU=0.50:0.95 | area= large | maxDets= 20 ] = 0.049

GengDavid commented 6 years ago

@Tiamo666 Thanks! I'm a little busy these days, I'll update the results and model soon.

mingloo commented 6 years ago

Hi @GengDavid @Tiamo666

I've used the commit 8e85af2453d68766ae0d196a4da3cb6605b6a4eb to train ResNet50+ 256x192 model with GT bbox input and default parameter setting from scratch when epoch is set to 32 and the overall result 70.8 as below shown is slightly worse than the reported one 71.2:

Average Precision  (AP) @[ IoU=0.50:0.95 | area=   all | maxDets= 20 ] = 0.708
Average Precision  (AP) @[ IoU=0.50      | area=   all | maxDets= 20 ] = 0.905
Average Precision  (AP) @[ IoU=0.75      | area=   all | maxDets= 20 ] = 0.782
Average Precision  (AP) @[ IoU=0.50:0.95 | area=medium | maxDets= 20 ] = 0.683
Average Precision  (AP) @[ IoU=0.50:0.95 | area= large | maxDets= 20 ] = 0.749
Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets= 20 ] = 0.740
Average Recall     (AR) @[ IoU=0.50      | area=   all | maxDets= 20 ] = 0.918
Average Recall     (AR) @[ IoU=0.75      | area=   all | maxDets= 20 ] = 0.804
Average Recall     (AR) @[ IoU=0.50:0.95 | area=medium | maxDets= 20 ] = 0.710
Average Recall     (AR) @[ IoU=0.50:0.95 | area= large | maxDets= 20 ] = 0.786

How many epochs do you set to achieve 71.2 for ResNet50+ 256x192 ?

As for ResNet50+ 384x288 model with GT bbox input and default parameter setting training from scratch, the epoch=32 result is slightly better than the reported 73.7 as follows:

Average Precision  (AP) @[ IoU=0.50:0.95 | area=   all | maxDets= 20 ] = 0.741
Average Precision  (AP) @[ IoU=0.50      | area=   all | maxDets= 20 ] = 0.925
Average Precision  (AP) @[ IoU=0.75      | area=   all | maxDets= 20 ] = 0.805
Average Precision  (AP) @[ IoU=0.50:0.95 | area=medium | maxDets= 20 ] = 0.706
Average Precision  (AP) @[ IoU=0.50:0.95 | area= large | maxDets= 20 ] = 0.795
Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets= 20 ] = 0.768
Average Recall     (AR) @[ IoU=0.50      | area=   all | maxDets= 20 ] = 0.932
Average Recall     (AR) @[ IoU=0.75      | area=   all | maxDets= 20 ] = 0.825
Average Recall     (AR) @[ IoU=0.50:0.95 | area=medium | maxDets= 20 ] = 0.730
Average Recall     (AR) @[ IoU=0.50:0.95 | area= large | maxDets= 20 ] = 0.826
GengDavid commented 6 years ago

@mkocabas Sorry about that I have not updated the results yet. 71.2 is the old result. It is strange that the results that after fixing bugs are lower than the results before. I'll update all the results this weekends, but I still do not figure out the reason. Maybe we need to adjust the parameter setting since this parameter setting is setting for the old codes.

mingloo commented 6 years ago

@GengDavid Now, the ResNet 50 + 256x192 with detection GT bboxes is slightly worse than the old result, but the ResNet 50 + 384x288 is slightly better than the old result.

GengDavid commented 6 years ago

Cool, so I think it is allowable to have some slight differences. And could you provide your pre-trained ResNet 50 + 384x288 with us? It would be great.

mingloo commented 6 years ago

@GengDavid Please see my comments https://github.com/GengDavid/pytorch-cpn/issues/3#issuecomment-424928303

GengDavid commented 6 years ago

Sorry, I don't clearly understand what you mean by referencing comment-424928303😳

mingloo commented 6 years ago

@GengDavid Sorry, I misunderstand your comment. The trained model for ResNet50+ 384x288 can be found at GoogleDrive.

GengDavid commented 6 years ago

Hi @Tiamo666 @mingloo I've updated all the pre-trained models and results.
Sorry for taking a long time to update. Thanks for your great work!

GengDavid commented 6 years ago

However, it is a little confusing that the CPN-101-384x288 model perform even worse than CPN-50-384x288.
@Tiamo666 Could you show me the parameter setting you used to fine-tune the model? Thanks!
Have a good National Day.

mingloo commented 6 years ago

@GengDavid @Tiamo666 Thanks for updating the result. I'll try to train CPN-ResNet101-384x288 from scratch on my side.

GengDavid commented 6 years ago

@mingloo Great! Thanks.

Tiamo666 commented 6 years ago

@GengDavid , Thanks a lot, I just come back from my holiday. I didn't change any other parameters, I just modified the learning rate scheduler with pytorch built-in package optim.lr_scheduler, here is my code:

fine tune

        for k, v in pretrained_dict.items():
            if k in ['module.global_net.upsamples.0.1.bias',
            'module.global_net.upsamples.1.1.bias',
            'module.global_net.upsamples.2.1.bias']:
                continue
            new_dict[k] = v
        model.load_state_dict(new_dict)

adjust lr rate

scheduler = lr_scheduler.MultiStepLR(optimizer, milestones = cfg.lr_dec_epoch, gamma=cfg.lr_gamma) for epoch in range(args.start_epoch, args.epochs):

lr = adjust_learning_rate(optimizer, epoch, cfg.lr_dec_epoch, cfg.lr_gamma)

    scheduler.step(epoch)
    lr = optimizer.state_dict()['param_groups'][0]['lr']
    print('\nEpoch: %d | LR: %.8f' % (epoch + 1, lr))

The following is part of my log.txt, I fine tuned from epoch32, and the total epoch is 35:

30.000000 0.000031 102.073177
31.000000 0.000016 101.399609
32.000000 0.000016 101.165480
33.000000 0.000016 101.801196
34.000000 0.000016 101.328027
35.000000 0.000016 101.059933