No such file or directory: 'skeleton_data/skeletons5_info.txt'

[dwdjsy89]

When running Train.ipynb

the cell:X0, Y0, clip_indices, classes = load_my_data('skeleton_data/skeletons5_info.txt')

it has an error "FileNotFoundErrorTraceback (most recent call last)

in () ----> 1 X0, Y0, clip_indices, classes = load_my_data('skeleton_data/skeletons5_info.txt') in load_my_data(filepath) 23 24 def load_my_data(filepath): ---> 25 with open(filepath, 'r') as f: 26 dataset = simplejson.load(f) 27 FileNotFoundError: [Errno 2] No such file or directory: 'skeleton_data/skeletons5_info.txt' " Where can I get the "skeletons5_info.txt"? I have downloaded "source_images3" but didn't see "skeletons5_info.txt" in it. [problem.pdf](https://github.com/felixchenfy/Realtime-Action-Recognition/files/3244673/problem.pdf)

[felixchenfy]

Hello, this "skeletons5_info.txt" is used for training the model. So you need to first download the dataset (images with a person in it), run a script to detect the person's skeleton, and then run another script to store skeleton results into this "skeletons5_info.txt".

The detailed steps are as follows (Just the same as the README.md):

4. How to run: Training

1). First, detect skeleton from training images and save result to txt:

$ python src/run_detector.py --source txtscript

• Input: Images are from "data/source_images3/". See this file for downloading.
• Output:
  (1) Skeleton positions of each image in "src/skeleton_data/skeletons5/" in txt format.
  (2) Images with drawn skeleton, saved in "src/skeleton_data/skeletons5_images/".

2). Second, put skeleton from multiple txt into a single txt file:

$ python src/scripts/skeletons_info_generator.py

• Input: "src/skeleton_data/skeletons5/".
• Output: "src/skeleton_data/skeletons5_info.txt"

3). Third, open jupyter notebook, and run this file: "src/Train.ipynb".
The trained model will be saved to "model/trained_classifier.pickle".

Now, it's ready to run the real-time demo from your web camera.