NVIDIA-AI-IOT / trt_pose

Real-time pose estimation accelerated with NVIDIA TensorRT
MIT License
957 stars 290 forks source link

I want to train mobilenet_v2 to use trt_pose #132

Open gj-park opened 3 years ago

gj-park commented 3 years ago

Hi, To accelerate your pose estimation, I tried training mobilenet_v2. So first, I made I copy densenet.py and changed word 'densenet' to 'mobilenet_v2' including what's in vi editor. and second, I wanted to use train.py (trt_pose/trt_pose/train.py). But there was an error message like that. How should I write to implement train.py ???????

(I typed): python3 train.py --model=mobilenet_v2 usage: train.py [-h] config train.py: error: the following arguments are required: config

sinuku commented 3 years ago

you need the json file(config) it's in 'trt_pose/tasks/human_pose/experiments/'