ai4bharat / include Goto Github PK

Is there any other inference script used for evaluating the video stream? Is there a way to create a video similar to the one displayed on the website?
I was able to recreate the code but it is currently working only on the test data.
Here is the error when I try to feed a long video with a sentence: -

Traceback (most recent call last): File "evaluate.py", line 136, in <module> process_video(path, save_dir) File "/home/parag/Sign Language/INCLUDE/generate_keypoints.py", line 90, in process_video pose_x, pose_y = process_pose_keypoints(pose_results) File "/home/parag/Sign Language/INCLUDE/generate_keypoints.py", line 39, in process_pose_keypoints pose_x, pose_y = process_landmarks(pose) File "/home/parag/Sign Language/INCLUDE/generate_keypoints.py", line 16, in process_landmarks for landmark in landmarks.landmark: AttributeError: 'NoneType' object has no attribute 'landmark'

I am getting the following error on training the include50 model after keypoints generation:
Traceback (most recent call last):
File "/home/cilans/Desktop/Mlmodels/INCLUDE/runner.py", line 87, in
train_nn.fit(args)
File "/home/cilans/Desktop/Mlmodels/INCLUDE/train_nn.py", line 225, in fit
val_loss, val_acc = validate(val_dataloader, model, device)
File "/home/cilans/.local/lib/python3.10/site-packages/torch/utils/_contextlib.py", line 115, in decorate_context
return func(*args, **kwargs)
File "/home/cilans/Desktop/Mlmodels/INCLUDE/train_nn.py", line 69, in validate
for batch in pbar:
File "/home/cilans/.local/lib/python3.10/site-packages/tqdm/std.py", line 1178, in iter
for obj in iterable:
File "/home/cilans/.local/lib/python3.10/site-packages/torch/utils/data/dataloader.py", line 631, in next
data = self._next_data()
File "/home/cilans/.local/lib/python3.10/site-packages/torch/utils/data/dataloader.py", line 1346, in _next_data
return self._process_data(data)
File "/home/cilans/.local/lib/python3.10/site-packages/torch/utils/data/dataloader.py", line 1372, in _process_data
data.reraise()
File "/home/cilans/.local/lib/python3.10/site-packages/torch/_utils.py", line 705, in reraise
raise exception
ValueError: Caught ValueError in DataLoader worker process 0.
Original Traceback (most recent call last):
File "/home/cilans/.local/lib/python3.10/site-packages/torch/utils/data/_utils/worker.py", line 308, in _worker_loop
data = fetcher.fetch(index) # type: ignore[possibly-undefined]
File "/home/cilans/.local/lib/python3.10/site-packages/torch/utils/data/_utils/fetch.py", line 51, in fetch
data = [self.dataset[idx] for idx in possibly_batched_index]
File "/home/cilans/.local/lib/python3.10/site-packages/torch/utils/data/_utils/fetch.py", line 51, in
data = [self.dataset[idx] for idx in possibly_batched_index]
File "/home/cilans/Desktop/Mlmodels/INCLUDE/dataset.py", line 113, in getitem
.reshape(-1, 50)
ValueError: cannot reshape array of size 3894 into shape (50)

Can someone please help.

I am getting a similar error on running the pretrained model in evaluate mode.

Hello, thank you for compiling this dataset and your work in general. I tried downloading the dataset from this link but it appears that the link for Train-Test split for INCLUDE and INCLUDE50 is broken. Its not available on Zenodo and the google drive link also seems to be broken. Can you please fix the broken link or provide one in this thread?

The data available on Zenodo is missing the following files.

Days_and_Time/Second (Number)/MVI_4650.MOV
Days_and_Time/Second (Number)/MVI_5506.MOV
Days_and_Time/Second (Number)/MVI_5505.MOV
Days_and_Time/Second (Number)/MVI_5503.MOV
Days_and_Time/Second (Number)/MVI_4647.MOV
Days_and_Time/Second (Number)/MVI_5504.MOV
Days_and_Time/Second (Number)/MVI_4649.MOV
Days_and_Time/Second (Number)/MVI_4648.MOV

Is there any way to get this data?