gsig / pyvideoresearch Goto Github PK

I have to use the TVSum50 dataset on a video summarization task. The original video uses a frame rate of 30 fps and each frame is assigned an importance score from 1-5. I have to downsample the video to 3 fps but am not understanding how that will affect the importance scores. Can anyone help me here, please?

Hi,

I'm fine-tuning the I3D model on my dataset based on the pretrained model aj_rgb_imagenet.pth. But the pretained model seems not able to be loaded. Here's the message from log.txt.

Missing key(s) in state_dict: "module.basenet.logits.conv3d.weight", "module.basenet.logits.conv3d.bias", "module.basenet.Conv3d_1a_7x7.conv3d.weight", "module.basenet.Conv3d_1a_7x7.bn.weight", "module.basenet.Conv3d_1a_7x7.bn.bias", "module.basenet.Conv3d_1a_7x7.bn.running_mean", "module.basenet.Conv3d_1a_7x7.bn.running_var", "module.basenet.Conv3d_2b_1x1.conv3d.weight", "module.basenet.Conv3d_2b_1x1.bn.weight", "module.basenet.Conv3d_2b_1x1.bn.bias", "module.basenet.Conv3d_2b_1x1.bn.running_mean"
...

and

size mismatch for module.basenet.Conv3d_1a_7x7.conv3d.weight: copying a param with shape torch.Size([128, 512, 1, 1, 1]) from checkpoint, the shape in current model is torch.Size([64, 3, 7, 7, 7]).
size mismatch for module.basenet.Conv3d_1a_7x7.bn.weight: copying a param with shape torch.Size([128]) from checkpoint, the shape in current model is torch.Size([64]).
size mismatch for module.basenet.Conv3d_1a_7x7.bn.bias: copying a param with shape torch.Size([128]) from checkpoint, the shape in current model is torch.Size([64]).
...

I don't think the model is loaded correctly so do you have any clue or did I miss something? Here's my experiment code:

root = 'MyRoot'
mode = 'rgb'
version = 'v11_13'
args = [
    '--name', name,  # name is filename
    '--print-freq', '1',
    '--dataset', 'mydataset',
    '--arch', 'aj_i3d',
    '--lr', '1e-3',  # set as default explicitly
    '--lr-decay-rate', '20',
    '--optimizer', 'adam',
    '--epochs', '10',
    '--batch-size', '5',
    '--weight-decay', '0.0000001',
    '--cache-dir', root + 'caches/' + version + '/' + mode + '/train/',
    '--data', root + 'dataset/',
    '--train-file', root + 'dataset/labels/train.json',
    '--val-file', root + 'dataset/labels/validation.json',
    '--label-file', root + 'dataset/labels/labels.json',
    '--pretrained',
    # '--evaluate',  # evaluate only
    '--dropout', '0.5',
    '--nclass', '2',
    '--resume', root + 'caches/' + version + '/' + mode + '/train/' + name + '/model.pth.tar;'
                root + 'pretrained_model/aj_rgb_imagenet.pth',
    '--workers', '2',
    '--replace-last-layer',
    '--metric', 'val_top1',
    '--metrics', 'top1_metric',
]

Any help is appreciated.

We need your help!

Hi!
I'm trying to train the nonlocal net on the Charades dataset and this repo is being of great help! However, I've run your script nonlocal_resnet50_3d_charades.py but it achieved results lower than expected. In the script says that you've reached 32% mAP but my best_model.txt got:

train_loss 0.06904963731537
train_top1 27.30711486067315
train_top5 86.0374044127445
val_loss 0.11540985137626932
val_top1 30.0
val_top5 94.5945945945946
video_task_CharadesmAP 0.17685393374913622
video_task_videotop1 38.16425120772947
video_task_videotop5 147.66505636070855

Is there something else I should be paying attention to run this script?

Also, the paper for this network (https://arxiv.org/pdf/1711.07971.pdf) reports 37.5% mAP. The reason for this gap in the results is known?

Hi~the top1 val error of your model is 64.15625, and the top1 val error of nonlocal resnet 50 in Xiaolong Wang's CVPR18 paper is 73.3, are there some differents between the two models in the training process?

Hey gsig, thanks for your share!!
However, when I try the baseline experiment - i3d_mask_rcnn_ava.py,
I got the following error:

RuntimeError: Assertion `THCTensor_(checkGPU)(state, 4, input, target, output, total_weight)' failed. Some of weight/gradient/input tensors are located on different GPUs. Please move them to a single one. at /opt/conda/conda-bld/pytorch-nightly_1553145032991/work/aten/src/THCUNN/generic/ClassNLLCriterion.cu:28

It looks like that the code do not Parallel properly.
Could you please help me fix this bug?

Thank you for sharing the code. It help a lot.
But there is some confusion about the data loading procedure in charades_video_tsn.py.
Here is the codes：
``

    n = self.data['datas'][index]['n']
    if shift is None:
        shift = np.random.randint(n-self.train_gap-2)
    else:
        shift = int(shift * (n-self.train_gap-2))
    ss = [shift] + [np.random.randint(n-self.train_gap-2)
                    for _ in range(self.segments-1)]

``
In this way, ss would be a list including indexes of images to be loaded. But I find the values in ss not regular, which seems not consistent with TSN.
Did I miss something? By the way, what's the difference among charades_tsn.py, charades_video.py and charades_video_xx.py?
Looking forward to your reply.

Hi,
Could someone point me where the box information is used in I3D AVA. I think the whole image is being used (bounding box is not used any where)? Am I correct?

Regards,
Srikanth

Hi, I have a question about group_size argument for your non local block.

In the original paper and AlexHe7's repo, there is no mention of such argument.

Can you explain bit more about its purpose?

Thank you.

One of the steps to train the network says to duplicate and edit one of the experiment files under exp/ with appropriate parameters. I'm not sure where exp/ and the experiment files being referred to are.

Hi~ Thanks for your work.

The default bs of non-local resnet50 on Charades is 5. But following this setting, we(with @icyzhang0923 ) got an mAP with 29.98, which is an obvious gap with the report result 32.08. So we change the bs and get the following results:

On account of your released model is on epoch47. So we evaluate epoch45 merely. Other settings are all consistent with default options and employed the pretrained model you provided. So, could you please tell us your bs setting and gpu number?

Thanks a lot.

We need your help!

i found baseline_exp/async_tf_i3d_charades.py cannot run directly.
so i modify line 81 in models/criteria/async_tf_criterion.py as follow
idtime = [] for i in range(len(meta)): idtime.append((meta[i]['id'], meta[i]['time']))

i was confused about line 105 in models/criteria/async_tf_criterion.py
loss += self.loss(torch.nn.Sigmoid()(a), target) * self.orig_loss
what's the self.orig_loss mean?

Hi,
Thanks for your awesome project! I have a question about a piece of code:
videos = glob.glob('videos/*.mp4')
This means the ffmpeg command only focus on '.mp4' format video. However, there are also '.mp4' but also '.webm' and '.mkv' which can not be covered by this code. Will it lose videos which have '.mkv' and '.webm' format? Am I right?

Best,

Hi, I tried to load checkpoint "aj_rgb_charades.pth" in i3d_ava.py.
But I have accounted size mismatch of model.
Please confirm if current model is matched to model of checkpoint.

=> loading checkpoint 'C:\Users*\PycharmProjects\pyvideoresearch\charades_pretrained\aj_rgb_charades.pth'
Error(s) in loading state_dict for MyDataParallel:
Missing key(s) in state_dict: "module.basenet.logits.conv3d.weight", "module.basenet.logits.conv3d.bias", "module.basenet.Conv3d_1a_7x7.conv3d.weight",
・・・
size mismatch for module.basenet.logits.conv3d.weight: copying a param with shape torch.Size([157, 1024, 1, 1, 1]) from checkpoint, the shape in current model is torch.Size([81, 1024, 1, 1, 1]).
size mismatch for module.basenet.logits.conv3d.bias: copying a param with shape torch.Size([157]) from checkpoint, the shape in current model is torch.Size([81]).
size mismatch for module.basenet.Conv3d_1a_7x7.conv3d.weight: copying a param with shape torch.Size([128, 512, 1, 1, 1]) from checkpoint, the shape in current model is torch.Size([64, 3, 7, 7, 7]).
size mismatch for module.basenet.Conv3d_1a_7x7.bn.weight: copying a param with shape torch.Size([128]) from checkpoint, the shape in current model is torch.Size([64]).
size mismatch for module.basenet.Conv3d_1a_7x7.bn.bias: copying a param with shape torch.Size([128]) from checkpoint, the shape in current model is torch.Size([64]).
size mismatch for module.basenet.Conv3d_1a_7x7.bn.running_mean: copying a param with shape torch.Size([128]) from checkpoint, the shape in current model is torch.Size([64]).
size mismatch for module.basenet.Conv3d_1a_7x7.bn.running_var: copying a param with shape torch.Size([128]) from checkpoint, the shape in current model is torch.Size([64]).
size mismatch for module.basenet.Conv3d_1a_7x7.bn.num_batches_tracked: copying a param with shape torch.Size([128, 512, 1, 1, 1]) from checkpoint, the shape in current model is torch.Size([]).
size mismatch for module.basenet.Conv3d_2b_1x1.conv3d.weight: copying a param with shape torch.Size([128]) from checkpoint, the shape in current model is torch.Size([64, 64, 1, 1, 1]).
size mismatch for module.basenet.Conv3d_2b_1x1.bn.weight: copying a param with shape torch.Size([128]) from checkpoint, the shape in current model is torch.Size([64]).
size mismatch for module.basenet.Conv3d_2b_1x1.bn.bias: copying a param with shape torch.Size([128]) from checkpoint, the shape in current model is torch.Size([64]).
size mismatch for module.basenet.Conv3d_2b_1x1.bn.running_mean: copying a param with shape torch.Size([128]) from checkpoint, the shape in current model is torch.Size([64]).
size mismatch for module.basenet.Conv3d_2b_1x1.bn.running_var: copying a param with shape torch.Size([256, 128, 3, 3, 3]) from checkpoint, the shape in current model is torch.Size([64]).
size mismatch for module.basenet.Conv3d_2c_3x3.conv3d.weight: copying a param with shape torch.Size([256]) from checkpoint, the shape in current model is torch.Size([192, 64, 3, 3, 3]).
size mismatch for module.basenet.Conv3d_2c_3x3.bn.weight: copying a param with shape torch.Size([256]) from checkpoint, the shape in current model is torch.Size([192]).
size mismatch for module.basenet.Conv3d_2c_3x3.bn.bias: copying a param with shape torch.Size([256]) from checkpoint, the shape in current model is torch.Size([192]).
size mismatch for module.basenet.Conv3d_2c_3x3.bn.running_mean: copying a param with shape torch.Size([24, 512, 1, 1, 1]) from checkpoint, the shape in current model is torch.Size([192]).
size mismatch for module.basenet.Conv3d_2c_3x3.bn.running_var: copying a param with shape torch.Size([24]) from checkpoint, the shape in current model is torch.Size([192]).
size mismatch for module.basenet.Mixed_3b.b0.conv3d.weight: copying a param with shape torch.Size([24]) from checkpoint, the shape in current model is torch.Size([64, 192, 1, 1, 1]).
size mismatch for module.basenet.Mixed_3b.b0.bn.weight: copying a param with shape torch.Size([24]) from checkpoint, the shape in current model is torch.Size([64]).
size mismatch for module.basenet.Mixed_3b.b0.bn.bias: copying a param with shape torch.Size([64, 24, 3, 3, 3]) from checkpoint, the shape in current model is torch.Size([64]).
size mismatch for module.basenet.Mixed_3b.b1a.conv3d.weight: copying a param with shape torch.Size([64]) from checkpoint, the shape in current model is torch.Size([96, 192, 1, 1, 1]).
size mismatch for module.basenet.Mixed_3b.b1a.bn.weight: copying a param with shape torch.Size([64, 512, 1, 1, 1]) from checkpoint, the shape in current model is torch.Size([96]).
size mismatch for module.basenet.Mixed_3b.b1a.bn.bias: copying a param with shape torch.Size([64]) from checkpoint, the shape in current model is torch.Size([96]).
size mismatch for module.basenet.Mixed_3b.b1a.bn.running_mean: copying a param with shape torch.Size([64]) from checkpoint, the shape in current model is torch.Size([96]).
size mismatch for module.basenet.Mixed_3b.b1a.bn.running_var: copying a param with shape torch.Size([64]) from checkpoint, the shape in current model is torch.Size([96]).
size mismatch for module.basenet.Mixed_3b.b1b.conv3d.weight: copying a param with shape torch.Size([112, 512, 1, 1, 1]) from checkpoint, the shape in current model is torch.Size([128, 96, 3, 3, 3]).
size mismatch for module.basenet.Mixed_3b.b1b.bn.weight: copying a param with shape torch.Size([112]) from checkpoint, the shape in current model is torch.Size([128]).
size mismatch for module.basenet.Mixed_3b.b1b.bn.bias: copying a param with shape torch.Size([112]) from checkpoint, the shape in current model is torch.Size([128]).
size mismatch for module.basenet.Mixed_3b.b1b.bn.running_mean: copying a param with shape torch.Size([112]) from checkpoint, the shape in current model is torch.Size([128]).
size mismatch for module.basenet.Mixed_3b.b1b.bn.running_var: copying a param with shape torch.Size([112]) from checkpoint, the shape in current model is torch.Size([128]).
size mismatch for module.basenet.Mixed_3b.b1b.bn.num_batches_tracked: copying a param with shape torch.Size([144, 512, 1, 1, 1]) from checkpoint, the shape in current model is torch.Size([]).
size mismatch for module.basenet.Mixed_3b.b2a.conv3d.weight: copying a param with shape torch.Size([144]) from checkpoint, the shape in current model is torch.Size([16, 192, 1, 1, 1]).
size mismatch for module.basenet.Mixed_3b.b2a.bn.weight: copying a param with shape torch.Size([144]) from checkpoint, the shape in current model is torch.Size([16]).
size mismatch for module.basenet.Mixed_3b.b2a.bn.bias: copying a param with shape torch.Size([144]) from checkpoint, the shape in current model is torch.Size([16]).
size mismatch for module.basenet.Mixed_3b.b2a.bn.running_mean: copying a param with shape torch.Size([144]) from checkpoint, the shape in current model is torch.Size([16]).
size mismatch for module.basenet.Mixed_3b.b2a.bn.running_var: copying a param with shape torch.Size([288, 144, 3, 3, 3]) from checkpoint, the shape in current model is torch.Size([16]).
size mismatch for module.basenet.Mixed_3b.b2b.conv3d.weight: copying a param with shape torch.Size([288]) from checkpoint, the shape in current model is torch.Size([32, 16, 3, 3, 3]).
size mismatch for module.basenet.Mixed_3b.b2b.bn.weight: copying a param with shape torch.Size([288]) from checkpoint, the shape in current model is torch.Size([32]).
size mismatch for module.basenet.Mixed_3b.b2b.bn.bias: copying a param with shape torch.Size([288]) from checkpoint, the shape in current model is torch.Size([32]).
size mismatch for module.basenet.Mixed_3b.b2b.bn.running_mean: copying a param with shape torch.Size([32, 512, 1, 1, 1]) from checkpoint, the shape in current model is torch.Size([32]).
size mismatch for module.basenet.Mixed_3b.b3b.conv3d.weight: copying a param with shape torch.Size([32]) from checkpoint, the shape in current model is torch.Size([32, 192, 1, 1, 1]).
size mismatch for module.basenet.Mixed_3b.b3b.bn.bias: copying a param with shape torch.Size([64, 32, 3, 3, 3]) from checkpoint, the shape in current model is torch.Size([32]).
size mismatch for module.basenet.Mixed_3b.b3b.bn.running_mean: copying a param with shape torch.Size([64]) from checkpoint, the shape in current model is torch.Size([32]).
size mismatch for module.basenet.Mixed_3b.b3b.bn.running_var: copying a param with shape torch.Size([64]) from checkpoint, the shape in current model is torch.Size([32]).
size mismatch for module.basenet.Mixed_3c.b0.conv3d.weight: copying a param with shape torch.Size([64]) from checkpoint, the shape in current model is torch.Size([128, 256, 1, 1, 1]).
size mismatch for module.basenet.Mixed_3c.b0.bn.weight: copying a param with shape torch.Size([64, 512, 1, 1, 1]) from checkpoint, the shape in current model is torch.Size([128]).
size mismatch for module.basenet.Mixed_3c.b0.bn.bias: copying a param with shape torch.Size([64]) from checkpoint, the shape in current model is torch.Size([128]).
size mismatch for module.basenet.Mixed_3c.b0.bn.running_mean: copying a param with shape torch.Size([64]) from checkpoint, the shape in current model is torch.Size([128]).
size mismatch for module.basenet.Mixed_3c.b0.bn.running_var: copying a param with shape torch.Size([64]) from checkpoint, the shape in current model is torch.Size([128]).
size mismatch for module.basenet.Mixed_3c.b1a.bn.num_batches_tracked: copying a param with shape torch.Size([128, 256, 1, 1, 1]) from checkpoint, the shape in current model is torch.Size([]).
size mismatch for module.basenet.Mixed_3c.b1b.conv3d.weight: copying a param with shape torch.Size([128]) from checkpoint, the shape in current model is torch.Size([192, 128, 3, 3, 3]).
size mismatch for module.basenet.Mixed_3c.b1b.bn.weight: copying a param with shape torch.Size([128]) from checkpoint, the shape in current model is torch.Size([192]).
size mismatch for module.basenet.Mixed_3c.b1b.bn.bias: copying a param with shape torch.Size([128]) from checkpoint, the shape in current model is torch.Size([192]).
size mismatch for module.basenet.Mixed_3c.b1b.bn.running_mean: copying a param with shape torch.Size([128]) from checkpoint, the shape in current model is torch.Size([192]).
size mismatch for module.basenet.Mixed_3c.b1b.bn.running_var: copying a param with shape torch.Size([192, 128, 3, 3, 3]) from checkpoint, the shape in current model is torch.Size([192]).
size mismatch for module.basenet.Mixed_3c.b2a.conv3d.weight: copying a param with shape torch.Size([192]) from checkpoint, the shape in current model is torch.Size([32, 256, 1, 1, 1]).
size mismatch for module.basenet.Mixed_3c.b2a.bn.weight: copying a param with shape torch.Size([192]) from checkpoint, the shape in current model is torch.Size([32]).
size mismatch for module.basenet.Mixed_3c.b2a.bn.bias: copying a param with shape torch.Size([192]) from checkpoint, the shape in current model is torch.Size([32]).
size mismatch for module.basenet.Mixed_3c.b2a.bn.running_mean: copying a param with shape torch.Size([32, 256, 1, 1, 1]) from checkpoint, the shape in current model is torch.Size([32]).
size mismatch for module.basenet.Mixed_3c.b2b.conv3d.weight: copying a param with shape torch.Size([32]) from checkpoint, the shape in current model is torch.Size([96, 32, 3, 3, 3]).
size mismatch for module.basenet.Mixed_3c.b2b.bn.weight: copying a param with shape torch.Size([32]) from checkpoint, the shape in current model is torch.Size([96]).
size mismatch for module.basenet.Mixed_3c.b2b.bn.bias: copying a param with shape torch.Size([96, 32, 3, 3, 3]) from checkpoint, the shape in current model is torch.Size([96]).
size mismatch for module.basenet.Mixed_3c.b3b.conv3d.weight: copying a param with shape torch.Size([96]) from checkpoint, the shape in current model is torch.Size([64, 256, 1, 1, 1]).
size mismatch for module.basenet.Mixed_3c.b3b.bn.weight: copying a param with shape torch.Size([64, 256, 1, 1, 1]) from checkpoint, the shape in current model is torch.Size([64]).
size mismatch for module.basenet.Mixed_4b.b0.conv3d.weight: copying a param with shape torch.Size([64, 192, 1, 1, 1]) from checkpoint, the shape in current model is torch.Size([192, 480, 1, 1, 1]).
size mismatch for module.basenet.Mixed_4b.b0.bn.weight: copying a param with shape torch.Size([64]) from checkpoint, the shape in current model is torch.Size([192]).
size mismatch for module.basenet.Mixed_4b.b0.bn.bias: copying a param with shape torch.Size([64]) from checkpoint, the shape in current model is torch.Size([192]).
size mismatch for module.basenet.Mixed_4b.b0.bn.running_mean: copying a param with shape torch.Size([64]) from checkpoint, the shape in current model is torch.Size([192]).
size mismatch for module.basenet.Mixed_4b.b0.bn.running_var: copying a param with shape torch.Size([64]) from checkpoint, the shape in current model is torch.Size([192]).
size mismatch for module.basenet.Mixed_4b.b0.bn.num_batches_tracked: copying a param with shape torch.Size([96, 192, 1, 1, 1]) from checkpoint, the shape in current model is torch.Size([]).
size mismatch for module.basenet.Mixed_4b.b1a.conv3d.weight: copying a param with shape torch.Size([96]) from checkpoint, the shape in current model is torch.Size([96, 480, 1, 1, 1]).
size mismatch for module.basenet.Mixed_4b.b1a.bn.running_var: copying a param with shape torch.Size([128, 96, 3, 3, 3]) from checkpoint, the shape in current model is torch.Size([96]).
size mismatch for module.basenet.Mixed_4b.b1b.conv3d.weight: copying a param with shape torch.Size([128]) from checkpoint, the shape in current model is torch.Size([208, 96, 3, 3, 3]).
size mismatch for module.basenet.Mixed_4b.b1b.bn.weight: copying a param with shape torch.Size([128]) from checkpoint, the shape in current model is torch.Size([208]).
size mismatch for module.basenet.Mixed_4b.b1b.bn.bias: copying a param with shape torch.Size([128]) from checkpoint, the shape in current model is torch.Size([208]).
size mismatch for module.basenet.Mixed_4b.b1b.bn.running_mean: copying a param with shape torch.Size([16, 192, 1, 1, 1]) from checkpoint, the shape in current model is torch.Size([208]).
size mismatch for module.basenet.Mixed_4b.b1b.bn.running_var: copying a param with shape torch.Size([16]) from checkpoint, the shape in current model is torch.Size([208]).
size mismatch for module.basenet.Mixed_4b.b2a.conv3d.weight: copying a param with shape torch.Size([16]) from checkpoint, the shape in current model is torch.Size([16, 480, 1, 1, 1]).
size mismatch for module.basenet.Mixed_4b.b2a.bn.bias: copying a param with shape torch.Size([32, 16, 3, 3, 3]) from checkpoint, the shape in current model is torch.Size([16]).
size mismatch for module.basenet.Mixed_4b.b2a.bn.running_mean: copying a param with shape torch.Size([32]) from checkpoint, the shape in current model is torch.Size([16]).
size mismatch for module.basenet.Mixed_4b.b2a.bn.running_var: copying a param with shape torch.Size([32]) from checkpoint, the shape in current model is torch.Size([16]).
size mismatch for module.basenet.Mixed_4b.b2b.conv3d.weight: copying a param with shape torch.Size([32]) from checkpoint, the shape in current model is torch.Size([48, 16, 3, 3, 3]).
size mismatch for module.basenet.Mixed_4b.b2b.bn.weight: copying a param with shape torch.Size([32, 192, 1, 1, 1]) from checkpoint, the shape in current model is torch.Size([48]).
size mismatch for module.basenet.Mixed_4b.b2b.bn.bias: copying a param with shape torch.Size([32]) from checkpoint, the shape in current model is torch.Size([48]).
size mismatch for module.basenet.Mixed_4b.b2b.bn.running_mean: copying a param with shape torch.Size([32]) from checkpoint, the shape in current model is torch.Size([48]).
size mismatch for module.basenet.Mixed_4b.b2b.bn.running_var: copying a param with shape torch.Size([32]) from checkpoint, the shape in current model is torch.Size([48]).
size mismatch for module.basenet.Mixed_4b.b3b.conv3d.weight: copying a param with shape torch.Size([384, 832, 1, 1, 1]) from checkpoint, the shape in current model is torch.Size([64, 480, 1, 1, 1]).
size mismatch for module.basenet.Mixed_4b.b3b.bn.weight: copying a param with shape torch.Size([384]) from checkpoint, the shape in current model is torch.Size([64]).
size mismatch for module.basenet.Mixed_4b.b3b.bn.bias: copying a param with shape torch.Size([384]) from checkpoint, the shape in current model is torch.Size([64]).
size mismatch for module.basenet.Mixed_4b.b3b.bn.running_mean: copying a param with shape torch.Size([384]) from checkpoint, the shape in current model is torch.Size([64]).
size mismatch for module.basenet.Mixed_4b.b3b.bn.running_var: copying a param with shape torch.Size([384]) from checkpoint, the shape in current model is torch.Size([64]).
size mismatch for module.basenet.Mixed_4b.b3b.bn.num_batches_tracked: copying a param with shape torch.Size([192, 832, 1, 1, 1]) from checkpoint, the shape in current model is torch.Size([]).
size mismatch for module.basenet.Mixed_4c.b0.conv3d.weight: copying a param with shape torch.Size([192]) from checkpoint, the shape in current model is torch.Size([160, 512, 1, 1, 1]).
size mismatch for module.basenet.Mixed_4c.b0.bn.weight: copying a param with shape torch.Size([192]) from checkpoint, the shape in current model is torch.Size([160]).
size mismatch for module.basenet.Mixed_4c.b0.bn.bias: copying a param with shape torch.Size([192]) from checkpoint, the shape in current model is torch.Size([160]).
size mismatch for module.basenet.Mixed_4c.b0.bn.running_mean: copying a param with shape torch.Size([192]) from checkpoint, the shape in current model is torch.Size([160]).
size mismatch for module.basenet.Mixed_4c.b0.bn.running_var: copying a param with shape torch.Size([384, 192, 3, 3, 3]) from checkpoint, the shape in current model is torch.Size([160]).
size mismatch for module.basenet.Mixed_4c.b1a.conv3d.weight: copying a param with shape torch.Size([384]) from checkpoint, the shape in current model is torch.Size([112, 512, 1, 1, 1]).
size mismatch for module.basenet.Mixed_4c.b1a.bn.weight: copying a param with shape torch.Size([384]) from checkpoint, the shape in current model is torch.Size([112]).
size mismatch for module.basenet.Mixed_4c.b1a.bn.bias: copying a param with shape torch.Size([384]) from checkpoint, the shape in current model is torch.Size([112]).
size mismatch for module.basenet.Mixed_4c.b1a.bn.running_mean: copying a param with shape torch.Size([48, 832, 1, 1, 1]) from checkpoint, the shape in current model is torch.Size([112]).
size mismatch for module.basenet.Mixed_4c.b1a.bn.running_var: copying a param with shape torch.Size([48]) from checkpoint, the shape in current model is torch.Size([112]).
size mismatch for module.basenet.Mixed_4c.b1b.conv3d.weight: copying a param with shape torch.Size([48]) from checkpoint, the shape in current model is torch.Size([224, 112, 3, 3, 3]).
size mismatch for module.basenet.Mixed_4c.b1b.bn.weight: copying a param with shape torch.Size([48]) from checkpoint, the shape in current model is torch.Size([224]).
size mismatch for module.basenet.Mixed_4c.b1b.bn.bias: copying a param with shape torch.Size([128, 48, 3, 3, 3]) from checkpoint, the shape in current model is torch.Size([224]).
size mismatch for module.basenet.Mixed_4c.b1b.bn.running_mean: copying a param with shape torch.Size([128]) from checkpoint, the shape in current model is torch.Size([224]).
size mismatch for module.basenet.Mixed_4c.b1b.bn.running_var: copying a param with shape torch.Size([128]) from checkpoint, the shape in current model is torch.Size([224]).
size mismatch for module.basenet.Mixed_4c.b2a.conv3d.weight: copying a param with shape torch.Size([128]) from checkpoint, the shape in current model is torch.Size([24, 512, 1, 1, 1]).
size mismatch for module.basenet.Mixed_4c.b2a.bn.weight: copying a param with shape torch.Size([128, 832, 1, 1, 1]) from checkpoint, the shape in current model is torch.Size([24]).
size mismatch for module.basenet.Mixed_4c.b2a.bn.bias: copying a param with shape torch.Size([128]) from checkpoint, the shape in current model is torch.Size([24]).
size mismatch for module.basenet.Mixed_4c.b2a.bn.running_mean: copying a param with shape torch.Size([128]) from checkpoint, the shape in current model is torch.Size([24]).
size mismatch for module.basenet.Mixed_4c.b2a.bn.running_var: copying a param with shape torch.Size([128]) from checkpoint, the shape in current model is torch.Size([24]).
size mismatch for module.basenet.Mixed_4c.b2b.conv3d.weight: copying a param with shape torch.Size([256, 832, 1, 1, 1]) from checkpoint, the shape in current model is torch.Size([64, 24, 3, 3, 3]).
size mismatch for module.basenet.Mixed_4c.b2b.bn.weight: copying a param with shape torch.Size([256]) from checkpoint, the shape in current model is torch.Size([64]).
size mismatch for module.basenet.Mixed_4c.b2b.bn.bias: copying a param with shape torch.Size([256]) from checkpoint, the shape in current model is torch.Size([64]).
size mismatch for module.basenet.Mixed_4c.b2b.bn.running_mean: copying a param with shape torch.Size([256]) from checkpoint, the shape in current model is torch.Size([64]).
size mismatch for module.basenet.Mixed_4c.b2b.bn.running_var: copying a param with shape torch.Size([256]) from checkpoint, the shape in current model is torch.Size([64]).
size mismatch for module.basenet.Mixed_4c.b2b.bn.num_batches_tracked: copying a param with shape torch.Size([160, 832, 1, 1, 1]) from checkpoint, the shape in current model is torch.Size([]).
size mismatch for module.basenet.Mixed_4c.b3b.conv3d.weight: copying a param with shape torch.Size([160]) from checkpoint, the shape in current model is torch.Size([64, 512, 1, 1, 1]).
size mismatch for module.basenet.Mixed_4c.b3b.bn.weight: copying a param with shape torch.Size([160]) from checkpoint, the shape in current model is torch.Size([64]).
size mismatch for module.basenet.Mixed_4c.b3b.bn.bias: copying a param with shape torch.Size([160]) from checkpoint, the shape in current model is torch.Size([64]).
size mismatch for module.basenet.Mixed_4c.b3b.bn.running_mean: copying a param with shape torch.Size([160]) from checkpoint, the shape in current model is torch.Size([64]).
size mismatch for module.basenet.Mixed_4c.b3b.bn.running_var: copying a param with shape torch.Size([320, 160, 3, 3, 3]) from checkpoint, the shape in current model is torch.Size([64]).
size mismatch for module.basenet.Mixed_4d.b0.conv3d.weight: copying a param with shape torch.Size([320]) from checkpoint, the shape in current model is torch.Size([128, 512, 1, 1, 1]).
size mismatch for module.basenet.Mixed_4d.b0.bn.weight: copying a param with shape torch.Size([320]) from checkpoint, the shape in current model is torch.Size([128]).
size mismatch for module.basenet.Mixed_4d.b0.bn.bias: copying a param with shape torch.Size([320]) from checkpoint, the shape in current model is torch.Size([128]).
size mismatch for module.basenet.Mixed_4d.b0.bn.running_mean: copying a param with shape torch.Size([32, 832, 1, 1, 1]) from checkpoint, the shape in current model is torch.Size([128]).
size mismatch for module.basenet.Mixed_4d.b0.bn.running_var: copying a param with shape torch.Size([32]) from checkpoint, the shape in current model is torch.Size([128]).
size mismatch for module.basenet.Mixed_4d.b1a.conv3d.weight: copying a param with shape torch.Size([32]) from checkpoint, the shape in current model is torch.Size([128, 512, 1, 1, 1]).
size mismatch for module.basenet.Mixed_4d.b1a.bn.weight: copying a param with shape torch.Size([32]) from checkpoint, the shape in current model is torch.Size([128]).
size mismatch for module.basenet.Mixed_4d.b1a.bn.bias: copying a param with shape torch.Size([128, 32, 3, 3, 3]) from checkpoint, the shape in current model is torch.Size([128]).
size mismatch for module.basenet.Mixed_4d.b1b.conv3d.weight: copying a param with shape torch.Size([128]) from checkpoint, the shape in current model is torch.Size([256, 128, 3, 3, 3]).
size mismatch for module.basenet.Mixed_4d.b1b.bn.weight: copying a param with shape torch.Size([128, 832, 1, 1, 1]) from checkpoint, the shape in current model is torch.Size([256]).
size mismatch for module.basenet.Mixed_4d.b1b.bn.bias: copying a param with shape torch.Size([128]) from checkpoint, the shape in current model is torch.Size([256]).
size mismatch for module.basenet.Mixed_4d.b1b.bn.running_mean: copying a param with shape torch.Size([128]) from checkpoint, the shape in current model is torch.Size([256]).
size mismatch for module.basenet.Mixed_4d.b1b.bn.running_var: copying a param with shape torch.Size([128]) from checkpoint, the shape in current model is torch.Size([256]).
size mismatch for module.basenet.Mixed_4d.b2a.conv3d.weight: copying a param with shape torch.Size([192, 480, 1, 1, 1]) from checkpoint, the shape in current model is torch.Size([24, 512, 1, 1, 1]).
size mismatch for module.basenet.Mixed_4d.b2a.bn.weight: copying a param with shape torch.Size([192]) from checkpoint, the shape in current model is torch.Size([24]).
size mismatch for module.basenet.Mixed_4d.b2a.bn.bias: copying a param with shape torch.Size([192]) from checkpoint, the shape in current model is torch.Size([24]).
size mismatch for module.basenet.Mixed_4d.b2a.bn.running_mean: copying a param with shape torch.Size([192]) from checkpoint, the shape in current model is torch.Size([24]).
size mismatch for module.basenet.Mixed_4d.b2a.bn.running_var: copying a param with shape torch.Size([192]) from checkpoint, the shape in current model is torch.Size([24]).
size mismatch for module.basenet.Mixed_4d.b2a.bn.num_batches_tracked: copying a param with shape torch.Size([96, 480, 1, 1, 1]) from checkpoint, the shape in current model is torch.Size([]).
size mismatch for module.basenet.Mixed_4d.b2b.conv3d.weight: copying a param with shape torch.Size([96]) from checkpoint, the shape in current model is torch.Size([64, 24, 3, 3, 3]).
size mismatch for module.basenet.Mixed_4d.b2b.bn.weight: copying a param with shape torch.Size([96]) from checkpoint, the shape in current model is torch.Size([64]).
size mismatch for module.basenet.Mixed_4d.b2b.bn.bias: copying a param with shape torch.Size([96]) from checkpoint, the shape in current model is torch.Size([64]).
size mismatch for module.basenet.Mixed_4d.b2b.bn.running_mean: copying a param with shape torch.Size([96]) from checkpoint, the shape in current model is torch.Size([64]).
size mismatch for module.basenet.Mixed_4d.b2b.bn.running_var: copying a param with shape torch.Size([208, 96, 3, 3, 3]) from checkpoint, the shape in current model is torch.Size([64]).
size mismatch for module.basenet.Mixed_4d.b3b.conv3d.weight: copying a param with shape torch.Size([208]) from checkpoint, the shape in current model is torch.Size([64, 512, 1, 1, 1]).
size mismatch for module.basenet.Mixed_4d.b3b.bn.weight: copying a param with shape torch.Size([208]) from checkpoint, the shape in current model is torch.Size([64]).
size mismatch for module.basenet.Mixed_4d.b3b.bn.bias: copying a param with shape torch.Size([208]) from checkpoint, the shape in current model is torch.Size([64]).
size mismatch for module.basenet.Mixed_4d.b3b.bn.running_mean: copying a param with shape torch.Size([16, 480, 1, 1, 1]) from checkpoint, the shape in current model is torch.Size([64]).
size mismatch for module.basenet.Mixed_4d.b3b.bn.running_var: copying a param with shape torch.Size([16]) from checkpoint, the shape in current model is torch.Size([64]).
size mismatch for module.basenet.Mixed_4e.b0.conv3d.weight: copying a param with shape torch.Size([16]) from checkpoint, the shape in current model is torch.Size([112, 512, 1, 1, 1]).
size mismatch for module.basenet.Mixed_4e.b0.bn.weight: copying a param with shape torch.Size([16]) from checkpoint, the shape in current model is torch.Size([112]).
size mismatch for module.basenet.Mixed_4e.b0.bn.bias: copying a param with shape torch.Size([48, 16, 3, 3, 3]) from checkpoint, the shape in current model is torch.Size([112]).
size mismatch for module.basenet.Mixed_4e.b0.bn.running_mean: copying a param with shape torch.Size([48]) from checkpoint, the shape in current model is torch.Size([112]).
size mismatch for module.basenet.Mixed_4e.b0.bn.running_var: copying a param with shape torch.Size([48]) from checkpoint, the shape in current model is torch.Size([112]).
size mismatch for module.basenet.Mixed_4e.b1a.conv3d.weight: copying a param with shape torch.Size([48]) from checkpoint, the shape in current model is torch.Size([144, 512, 1, 1, 1]).
size mismatch for module.basenet.Mixed_4e.b1a.bn.weight: copying a param with shape torch.Size([64, 480, 1, 1, 1]) from checkpoint, the shape in current model is torch.Size([144]).
size mismatch for module.basenet.Mixed_4e.b1a.bn.bias: copying a param with shape torch.Size([64]) from checkpoint, the shape in current model is torch.Size([144]).
size mismatch for module.basenet.Mixed_4e.b1a.bn.running_mean: copying a param with shape torch.Size([64]) from checkpoint, the shape in current model is torch.Size([144]).
size mismatch for module.basenet.Mixed_4e.b1a.bn.running_var: copying a param with shape torch.Size([64]) from checkpoint, the shape in current model is torch.Size([144]).
size mismatch for module.basenet.Mixed_4e.b1b.conv3d.weight: copying a param with shape torch.Size([160, 512, 1, 1, 1]) from checkpoint, the shape in current model is torch.Size([288, 144, 3, 3, 3]).
size mismatch for module.basenet.Mixed_4e.b1b.bn.weight: copying a param with shape torch.Size([160]) from checkpoint, the shape in current model is torch.Size([288]).
size mismatch for module.basenet.Mixed_4e.b1b.bn.bias: copying a param with shape torch.Size([160]) from checkpoint, the shape in current model is torch.Size([288]).
size mismatch for module.basenet.Mixed_4e.b1b.bn.running_mean: copying a param with shape torch.Size([160]) from checkpoint, the shape in current model is torch.Size([288]).
size mismatch for module.basenet.Mixed_4e.b1b.bn.running_var: copying a param with shape torch.Size([160]) from checkpoint, the shape in current model is torch.Size([288]).
size mismatch for module.basenet.Mixed_4e.b1b.bn.num_batches_tracked: copying a param with shape torch.Size([112, 512, 1, 1, 1]) from checkpoint, the shape in current model is torch.Size([]).
size mismatch for module.basenet.Mixed_4e.b2a.conv3d.weight: copying a param with shape torch.Size([112]) from checkpoint, the shape in current model is torch.Size([32, 512, 1, 1, 1]).
size mismatch for module.basenet.Mixed_4e.b2a.bn.weight: copying a param with shape torch.Size([112]) from checkpoint, the shape in current model is torch.Size([32]).
size mismatch for module.basenet.Mixed_4e.b2a.bn.bias: copying a param with shape torch.Size([112]) from checkpoint, the shape in current model is torch.Size([32]).
size mismatch for module.basenet.Mixed_4e.b2a.bn.running_mean: copying a param with shape torch.Size([112]) from checkpoint, the shape in current model is torch.Size([32]).
size mismatch for module.basenet.Mixed_4e.b2a.bn.running_var: copying a param with shape torch.Size([224, 112, 3, 3, 3]) from checkpoint, the shape in current model is torch.Size([32]).
size mismatch for module.basenet.Mixed_4e.b2b.conv3d.weight: copying a param with shape torch.Size([224]) from checkpoint, the shape in current model is torch.Size([64, 32, 3, 3, 3]).
size mismatch for module.basenet.Mixed_4e.b2b.bn.weight: copying a param with shape torch.Size([224]) from checkpoint, the shape in current model is torch.Size([64]).
size mismatch for module.basenet.Mixed_4e.b2b.bn.bias: copying a param with shape torch.Size([224]) from checkpoint, the shape in current model is torch.Size([64]).
size mismatch for module.basenet.Mixed_4e.b2b.bn.running_mean: copying a param with shape torch.Size([24, 512, 1, 1, 1]) from checkpoint, the shape in current model is torch.Size([64]).
size mismatch for module.basenet.Mixed_4e.b2b.bn.running_var: copying a param with shape torch.Size([24]) from checkpoint, the shape in current model is torch.Size([64]).
size mismatch for module.basenet.Mixed_4e.b3b.conv3d.weight: copying a param with shape torch.Size([24]) from checkpoint, the shape in current model is torch.Size([64, 512, 1, 1, 1]).
size mismatch for module.basenet.Mixed_4e.b3b.bn.weight: copying a param with shape torch.Size([24]) from checkpoint, the shape in current model is torch.Size([64]).
size mismatch for module.basenet.Mixed_4e.b3b.bn.bias: copying a param with shape torch.Size([64, 24, 3, 3, 3]) from checkpoint, the shape in current model is torch.Size([64]).
size mismatch for module.basenet.Mixed_4f.b0.conv3d.weight: copying a param with shape torch.Size([64]) from checkpoint, the shape in current model is torch.Size([256, 528, 1, 1, 1]).
size mismatch for module.basenet.Mixed_4f.b0.bn.weight: copying a param with shape torch.Size([64, 512, 1, 1, 1]) from checkpoint, the shape in current model is torch.Size([256]).
size mismatch for module.basenet.Mixed_4f.b0.bn.bias: copying a param with shape torch.Size([64]) from checkpoint, the shape in current model is torch.Size([256]).
size mismatch for module.basenet.Mixed_4f.b0.bn.running_mean: copying a param with shape torch.Size([64]) from checkpoint, the shape in current model is torch.Size([256]).
size mismatch for module.basenet.Mixed_4f.b0.bn.running_var: copying a param with shape torch.Size([64]) from checkpoint, the shape in current model is torch.Size([256]).
size mismatch for module.basenet.Mixed_4f.b1a.conv3d.weight: copying a param with shape torch.Size([64, 64, 1, 1, 1]) from checkpoint, the shape in current model is torch.Size([160, 528, 1, 1, 1]).
size mismatch for module.basenet.Mixed_4f.b1a.bn.weight: copying a param with shape torch.Size([64]) from checkpoint, the shape in current model is torch.Size([160]).
size mismatch for module.basenet.Mixed_4f.b1a.bn.bias: copying a param with shape torch.Size([64]) from checkpoint, the shape in current model is torch.Size([160]).
size mismatch for module.basenet.Mixed_4f.b1a.bn.running_mean: copying a param with shape torch.Size([64]) from checkpoint, the shape in current model is torch.Size([160]).
size mismatch for module.basenet.Mixed_4f.b1a.bn.running_var: copying a param with shape torch.Size([64]) from checkpoint, the shape in current model is torch.Size([160]).
size mismatch for module.basenet.Mixed_4f.b1a.bn.num_batches_tracked: copying a param with shape torch.Size([192, 64, 3, 3, 3]) from checkpoint, the shape in current model is torch.Size([]).
size mismatch for module.basenet.Mixed_4f.b1b.conv3d.weight: copying a param with shape torch.Size([192]) from checkpoint, the shape in current model is torch.Size([320, 160, 3, 3, 3]).
size mismatch for module.basenet.Mixed_4f.b1b.bn.weight: copying a param with shape torch.Size([192]) from checkpoint, the shape in current model is torch.Size([320]).
size mismatch for module.basenet.Mixed_4f.b1b.bn.bias: copying a param with shape torch.Size([192]) from checkpoint, the shape in current model is torch.Size([320]).
size mismatch for module.basenet.Mixed_4f.b1b.bn.running_mean: copying a param with shape torch.Size([192]) from checkpoint, the shape in current model is torch.Size([320]).
size mismatch for module.basenet.Mixed_4f.b1b.bn.running_var: copying a param with shape torch.Size([64, 3, 7, 7, 7]) from checkpoint, the shape in current model is torch.Size([320]).
size mismatch for module.basenet.Mixed_4f.b2a.conv3d.weight: copying a param with shape torch.Size([64]) from checkpoint, the shape in current model is torch.Size([32, 528, 1, 1, 1]).
size mismatch for module.basenet.Mixed_4f.b2a.bn.weight: copying a param with shape torch.Size([64]) from checkpoint, the shape in current model is torch.Size([32]).
size mismatch for module.basenet.Mixed_4f.b2a.bn.bias: copying a param with shape torch.Size([64]) from checkpoint, the shape in current model is torch.Size([32]).
size mismatch for module.basenet.Mixed_4f.b2a.bn.running_mean: copying a param with shape torch.Size([256, 528, 1, 1, 1]) from checkpoint, the shape in current model is torch.Size([32]).
size mismatch for module.basenet.Mixed_4f.b2a.bn.running_var: copying a param with shape torch.Size([256]) from checkpoint, the shape in current model is torch.Size([32]).
size mismatch for module.basenet.Mixed_4f.b2b.conv3d.weight: copying a param with shape torch.Size([256]) from checkpoint, the shape in current model is torch.Size([128, 32, 3, 3, 3]).
size mismatch for module.basenet.Mixed_4f.b2b.bn.weight: copying a param with shape torch.Size([256]) from checkpoint, the shape in current model is torch.Size([128]).
size mismatch for module.basenet.Mixed_4f.b2b.bn.bias: copying a param with shape torch.Size([160, 528, 1, 1, 1]) from checkpoint, the shape in current model is torch.Size([128]).
size mismatch for module.basenet.Mixed_4f.b2b.bn.running_mean: copying a param with shape torch.Size([160]) from checkpoint, the shape in current model is torch.Size([128]).
size mismatch for module.basenet.Mixed_4f.b2b.bn.running_var: copying a param with shape torch.Size([160]) from checkpoint, the shape in current model is torch.Size([128]).
size mismatch for module.basenet.Mixed_4f.b3b.conv3d.weight: copying a param with shape torch.Size([160]) from checkpoint, the shape in current model is torch.Size([128, 528, 1, 1, 1]).
size mismatch for module.basenet.Mixed_4f.b3b.bn.weight: copying a param with shape torch.Size([320, 160, 3, 3, 3]) from checkpoint, the shape in current model is torch.Size([128]).
size mismatch for module.basenet.Mixed_4f.b3b.bn.bias: copying a param with shape torch.Size([320]) from checkpoint, the shape in current model is torch.Size([128]).
size mismatch for module.basenet.Mixed_4f.b3b.bn.running_mean: copying a param with shape torch.Size([320]) from checkpoint, the shape in current model is torch.Size([128]).
size mismatch for module.basenet.Mixed_4f.b3b.bn.running_var: copying a param with shape torch.Size([320]) from checkpoint, the shape in current model is torch.Size([128]).
size mismatch for module.basenet.Mixed_5b.b0.conv3d.weight: copying a param with shape torch.Size([32, 528, 1, 1, 1]) from checkpoint, the shape in current model is torch.Size([256, 832, 1, 1, 1]).
size mismatch for module.basenet.Mixed_5b.b0.bn.weight: copying a param with shape torch.Size([32]) from checkpoint, the shape in current model is torch.Size([256]).
size mismatch for module.basenet.Mixed_5b.b0.bn.bias: copying a param with shape torch.Size([32]) from checkpoint, the shape in current model is torch.Size([256]).
size mismatch for module.basenet.Mixed_5b.b0.bn.running_mean: copying a param with shape torch.Size([32]) from checkpoint, the shape in current model is torch.Size([256]).
size mismatch for module.basenet.Mixed_5b.b0.bn.running_var: copying a param with shape torch.Size([32]) from checkpoint, the shape in current model is torch.Size([256]).
size mismatch for module.basenet.Mixed_5b.b0.bn.num_batches_tracked: copying a param with shape torch.Size([128, 32, 3, 3, 3]) from checkpoint, the shape in current model is torch.Size([]).
size mismatch for module.basenet.Mixed_5b.b1a.conv3d.weight: copying a param with shape torch.Size([128]) from checkpoint, the shape in current model is torch.Size([160, 832, 1, 1, 1]).
size mismatch for module.basenet.Mixed_5b.b1a.bn.weight: copying a param with shape torch.Size([128]) from checkpoint, the shape in current model is torch.Size([160]).
size mismatch for module.basenet.Mixed_5b.b1a.bn.bias: copying a param with shape torch.Size([128]) from checkpoint, the shape in current model is torch.Size([160]).
size mismatch for module.basenet.Mixed_5b.b1a.bn.running_mean: copying a param with shape torch.Size([128]) from checkpoint, the shape in current model is torch.Size([160]).
size mismatch for module.basenet.Mixed_5b.b1a.bn.running_var: copying a param with shape torch.Size([128, 528, 1, 1, 1]) from checkpoint, the shape in current model is torch.Size([160]).
size mismatch for module.basenet.Mixed_5b.b1b.conv3d.weight: copying a param with shape torch.Size([128]) from checkpoint, the shape in current model is torch.Size([320, 160, 3, 3, 3]).
size mismatch for module.basenet.Mixed_5b.b1b.bn.weight: copying a param with shape torch.Size([128]) from checkpoint, the shape in current model is torch.Size([320]).
size mismatch for module.basenet.Mixed_5b.b1b.bn.bias: copying a param with shape torch.Size([128]) from checkpoint, the shape in current model is torch.Size([320]).
loading partial state
ignoring state key for loading: logits.conv3d.weight
ignoring state key for loading: logits.conv3d.bias
=> loaded checkpoint 'C:\Users*\PycharmProjects\pyvideoresearch\charades_pretrained\aj_rgb_charades.pth' (just weights)

Hi. Thanks for this amazing repository.

I was wondering if you had any plans to load videos directly instead of first converting it to images. Having it in image format is very expensive from storage point of view. One way to directly load the video would be to use DALI (https://github.com/NVIDIA/DALI).

Hi, Thank you for a great repo!
I'd like to train AVA dataset for action recognition, but I'm confused how to train using code in this repo.
Please explain in detail.

Hi Gunnar,

I would like to reproduce the results of the model called I3D (with ResNet-101 backbone) reported in [1]. Which configuration should I use? If I just replace resnet50-3d with resnet101-3d, it seems to be not working.

[1] Videos as Space-Time Region Graphs, ECCV'18 [link]

Regards,
Avijit

When I run the TSN experiments on Charades, I meet a problem:

Traceback (most recent call last):
File "/media/icyzhang/SSD/PyVideoResearch-master/baseline_exp/temporal_segment_networks_charades.py", line 54, in
main()
File "/media/icyzhang/SSD/PyVideoResearch-master/main.py", line 106, in main
scores.update(trainer.train(train_loader, model, criterion, optimizer, epoch, metrics, args))
File "/media/icyzhang/SSD/PyVideoResearch-master/train.py", line 61, in train
output = model(input, meta)
File "/home/icyzhang/anaconda3/lib/python3.6/site-packages/torch/nn/modules/module.py", line 489, in call
result = self.forward(*input, **kwargs)
File "/home/icyzhang/anaconda3/lib/python3.6/site-packages/torch/nn/parallel/data_parallel.py", line 143, in forward
outputs = self.parallel_apply(replicas, inputs, kwargs)
File "/home/icyzhang/anaconda3/lib/python3.6/site-packages/torch/nn/parallel/data_parallel.py", line 153, in parallel_apply
return parallel_apply(replicas, inputs, kwargs, self.device_ids[:len(replicas)])
File "/home/icyzhang/anaconda3/lib/python3.6/site-packages/torch/nn/parallel/parallel_apply.py", line 83, in parallel_apply
raise output
File "/home/icyzhang/anaconda3/lib/python3.6/site-packages/torch/nn/parallel/parallel_apply.py", line 59, in _worker
output = module(*input, **kwargs)
TypeError: TSNBase2 object argument after ** must be a mapping, not tuple