NVIDIA-AI-IOT / tao-toolkit-triton-apps

Sample app code for deploying TAO Toolkit trained models to Triton
MIT License
84 stars 30 forks source link

(fix) Shape and bbox error #19

Open GMBarra opened 1 year ago

GMBarra commented 1 year ago

Add a small code to fix an error relate to pose_cls_dataset_convert.py when some times this condition while len(pose_sequences[object_id]) - frame_start >= sequence_length_min: was not met and the code add an extra None array to data_array. This generate a problem in the E2E inference in Pose Classification.

if data_array is not None:
    data_arrays.append(data_array)

The other small code I added was in plot_e2e_inference.py. I just add this to draw the bbox without the presence of the variable "bbox" in the files:

#Fix bbox, using minmax joints point
x_list,y_list = zip(*joints)
x_min = min(x_list)
x_max = max(x_list)
y_min = min(y_list)
y_max = max(y_list)
bbox = [x_min,y_min,x_max,y_max]