taesoo-kim/tcnactionrecognition

Skeleton based action recognition models with TCN variants for learning interpretable representation.

tcnactionrecognition's Issues

Pre-Trained model

Hi,
Thanks for sharing the code.
Is that possible to share the pre-trained model on NTU-RGBD dataset ?

Thanks,

Jue

missing skeletons file

Hi, could you also upload the file

/home-2/[email protected]/data/nturgbd/samples_with_missing_skeletons.txt

This file is missing and not sure how to get it

is this a bug?

Very good project!!!

I may find a little bug..
In
TCNActionRecognition/updates/train.py

line 157 : X = np.zeros((batch_size,max_len,feat_dim))

I think you mean is:
X = np.zeros((batch_size,max_len,feat_dim,1))

because in line 180:X[batch_count] = x.reshape(max_len,feat_dim,1)

bug in source code

Hi,
I replaced the update files with other ones. After running the train.Py I received the below error:

File "/mnt/homedata/ss/CodeProject/TCNActionRecognition/train.py", line 183, in nturgbd_train_datagen
X[batch_count] = x.reshape(max_len,feat_dim,1)
ValueError: could not broadcast input array from shape (300,150,1) into shape (300,150)

which with remove the paramet of 1 from reshape it resolved

ORIGINAL

  X[batch_count] =  x.reshape(max_len,feat_dim)

but now I have this error:
2018-04-16 10:34:10.496385: F ./tensorflow/core/util/mkl_util.h:1287] Non-OK-status: Status(error::Code::INVALID_ARGUMENT, "Unsupported data format") status: Invalid argument: Unsupported data format

I would be grateful if you guide me.
Best

Compatibility issue

hi

1. i had this error in train.py line 176 : x.setflags(write=1)
   " ValueError: cannot set WRITEABLE flag to True of this array"
   Exactly what versions of libraries are you using??
2. at file create_lmdb.py line 40
   "No such file : 'data_root+'Xy_train_%03d'%i+'.h5' '"
   thanx

cann't find the ".h5" file

Hi,
Thanks for sharing the code.
Im getting the following error. May I know where I can download this file?

No such file : 'data_root+'Xy_train_%03d'%i+'.h5' '

No file or directory

Im getting the following error. May I know where I can download this file?

No such file or directory: '/home/tk/dev/data/nturgbd/samples_with_missing_skeletons.txt'

visualize_filters.py

Hi,
I used the updated files and could train the data using train.py but I want use visualize_filters.py which there are several error (especially in a_conv1_filter dimension)due to it is seems this class is not compatible with the updated classes. is it possible to share the updated version of this class too.
Thanks a lot

dont understand the process of MEAN SUBTRACTION

Hi, kim, i dont really understand the process of MEAN SUBTRACTION in the train.py, especially those codes:

## THIS IS MEAN SUBTRACTION
      x = value.reshape((max_len,feat_dim))
      nonzeros = np.where(np.array([np.sum(x[i])>0 for i in range(0,x.shape[0])])==False)[0] 
      # i dont understand what nonzeros stand for ?
      if len(nonzeros) == 0:
        last_time = 0
      else:
        last_time = nonzeros[0]
      x.setflags(write=1)
      x[:last_time] = x[:last_time] - train_x_mean

would you please describe how it work? Thanks

求NTU RGB+D:A Large Scale Dataset for 3D Human Activity Analysis中PLSTM的源代码

你好：
我在https://github.com/FesianXu/PLSTM看见了您与作者的对话，不知您现在跑通代码没有，如果跑通了是否方便把整个工程给我一份呢？如果可以我的邮箱是[email protected]。将非常感谢您的帮助。期待您的回复

identify 'feat_dim'

Hi，I am now in Keras , and the variable' feat_dim' which I cant understand , could you describe it ? Thank you

How to get predicted values

Hello,
Can you please guide me how can I access to predicted scores in your code?
Best