opencap-core
opencap-core copied to clipboard
Main OpenCap processing pipeline
@suhlrich we might want to look into waht is happening with [this scaling](https://api.opencap.ai/sessions/c5c492d7-90af-417d-a80c-77f0a825ab07/), it looks bad: 
ffmpeg
Do we need ffmpeg in docker since we have it conda? It is a nightmare to build the images, sometimes it works sometimes it does not. Pinging you @suhlrich ChatGPT...
try 1x736_2scales if neutral.
Add an option in advanced settings Step 4.
Test [examples ](https://github.com/open-mmlab/mmpose/blob/master/demo/docs/2d_wholebody_pose_demo.md#speed-up-inference) 1. set flip_test=False in [pose_hrnet_w48_dark+](https://github.com/stanfordnmbl/opencap-core/blob/main/mmpose/hrnet_w48_coco_wholebody_384x288_dark_plus.py#L79). 2. set post_process='default' in [pose_hrnet_w48_dark+](https://github.com/stanfordnmbl/opencap-core/blob/main/mmpose/hrnet_w48_coco_wholebody_384x288_dark_plus.py#L80). 3. use faster human bounding box detector, see [MMDetection](https://mmdetection.readthedocs.io/en/latest/model_zoo.html).
See example w/ Transformer
investigate what went wrong with calibration here: https://api.opencap.ai/sessions/27a9972c-5705-4306-b89b-720a87acd0e7/
Bumps [scipy](https://github.com/scipy/scipy) from 1.9.0 to 1.10.0. Release notes Sourced from scipy's releases. SciPy 1.10.0 Release Notes SciPy 1.10.0 is the culmination of 6 months of hard work. It contains many...