#############SETTING UP THE ENVIRONMENTS,DOWNLOADING STUFFS############# 1.Ensure you are within /REPO directory

2.Run the command !pip install -r requirements.txt

#.If you are having any issue with installing torch and/or torchvisoin may be you'll like to do 'conda install pytorch torchvision -c pytorch' seperately to install torch

#############ARRANGING THE TO-BE-TESTED DATA#############

3.Put the images in /REPO/data/images folder

[the code isn't ready yet to handle videos by itself. Please transform the videos into imageframes and populate the image folder. This code snippet can be used]

#video frame capturer
def FrameCapture(video_name):
	vidObj = cv2.VideoCapture('/REPO/data/videos/'+video_name)
	count = 0
	success = 1
	while success:
  			success, image = vidObj.read()
  			cv2.imwrite("/REPO/data/images/frame%d.jpg" % count, image)
  			count += 1

#############GENERATING OUTPUT############# 3.Open the run_it.ipynb notebook and run the cells. It will show the predictions in the console and will also save the predicted information in a csv file named output.csv in the folder ~/Bosch/REPO/output

The code has been running fine on our local machines, but we fear there can be errors in file path with a change in the running environment. the required file paths may need to be modified if necessary#####

THE TEAM: AKASH GUPTA KUSHAGRA BHUSHAN JANHAVI SINGH SARJITA PATRA BHAVESH TANGIRALA SAYAN BISWAS

Weights need to be downloaded seperately