Fetching metadata from the HF Docker repository...
-
Update handlers/frame_handler_yolo.py
-
Update handlers/frame_handler_yolo.py
-
Update handlers/frame_handler_yolo.py
-
Upload football.mp4
-
Delete input_data/football.mp4
-
Upload football.mp4
-
Delete input_data/football.mp4
-
Rename input_data/football.mp4.mov to input_data/football.mp4
-
Upload football.mp4.mov
-
Delete input_data/football.mp4
-
Upload football.mp4
-
Delete input_data/football.mp4
-
Upload football.mp4
-
Delete input_data/football.mp4
-
Update app.py
-
Update app.py
-
Update app.py
-
Upload football.mp4
-
Delete input_data/football.mp4
-
revert code app.py as the approach to use GPU as needed doesnot work
-
refactor code to only use GPU as needed once deploying to HF
-
go back to 3 step appoach, increase model confidence, use ffmpeg
-
change duration to 300
-
implement new approach, ignore step 2: extracting key frames
-
revise the logic to use ultilize orignal fps
-
upload a new test file
-
Delete input_data/football.mp4
-
add description
-
add logic to auto load input file for better testing experience
-
Upload football.mp4
-
remove unnecessary dependencies
-
change video encoder to mp4
-
fix input video preview issue
-
fix video encoding error and improve layout
-
add imageio[ffmpeg] to requirements.txt
-
increase GPU duration to 600
-
fix unbound local error
-
improve performance by reducing file I/O
-
add msg to debug
-
add counter to frame handler
-
fix code caussing gpu issue
-
remove version from requirements.txt
-
fix build error
-
inital commit
-
initial commit
