Model fits the upper boddy well, but fails completly for lower body parts! Is this due to the wrong choice of the models?

[eulerleibniz]

I am testing your model with the following command:

python demo.py --gpu 0 --img_path input.jpg --output_folder output

I have the following output (Seems perfectly fine, just a warning):

Using GPU: 0
/home/user_name/miniconda3/envs/myenv/lib/python3.8/site-packages/mmcv/init.py:20: UserWarning: On January 1, 2023, MMCV will release v2.0.0, in which it will remove components related to the training process and add a data transformation module. In addition, it will rename the package names mmcv to mmcv-lite and mmcv-full to mmcv. See https://github.com/open-mmlab/mmcv/blob/master/docs/en/compatibility.md for more details.
warnings.warn(
06-13 04:22:40 Load checkpoint from ../pretrained_models/osx_l.pth.tar
06-13 04:22:40 Creating graph...
Load checkpoint from ../pretrained_models/osx_l.pth.tar
Using cache found in /home/user_name/.cache/torch/hub/ultralytics_yolov5_master
YOLOv5 🚀 2024-6-13 Python-3.8.19 torch-1.11.0+cu113 CUDA:0 (NVIDIA GeForce GTX 1070 Ti, 8111MiB)
Fusing layers...
YOLOv5s summary: 213 layers, 7225885 parameters, 0 gradients
Adding AutoShape...

Problem is that the upper body is well approximated but lower body is completely off. Specially in my single person images where the person is laying down or sitting with their feet offset from their body I get very awful results. (I can share images with you, but not doing it cause not sure of their copyright)

Am I using an incorrect model checkpoint? (shows osx_l.pth.tar). Am I using incorrect key-point detector (shows YOLOV5). Is your model trained on upper body only? How to change the models if there are options?