zhanghongrui

why set layer_index=28, filter_index=10? thanks?

There was a problem with training.

2

multigpu_train.py, The program is displayed as:"Generator use 10 batches for buffering, this may take a while, you can tune this yourself. 0 training images in /data/ocr/icdar2015/" Then it stopped moving.

sknet50

2

I downloaded sknet50, but failed to restore the model. Is there any code for sknet50(pytorch version)？ thanks error: self.base.load_state_dict(torch.load(model_path))

layer1.0.conv2.weight

1

why not layer1.0.conv2.weight in bottle_block_flag of get_small_model.py? only exist con1 and con3?

hi, python head_detection_demo.py have error

1

python head_detection_demo.py --img_path /home/fwq/pic/1.png --model_path /home/fwq/zhanghongruils/head/fchd/model/head_detector_final info: the python code for non_maximum_suppression is about 2x slow It is strongly recommended to build cython code: `cd model/utils/nms/; python3 build.py build_ext --inplace...

libtorch or tensorrt

12

hi,thank your code. Can this model be converted to Library or tensorRT? hope your relpy!

Implications for other tasks,For example,Is it valid for Person Reid? thanks your code

你好，谢谢你的代码 裁剪后的模型没有微调恢复一定的精度吗？

Baseline

1

Thanks for the UnrealStereo4K dataset，and What is its baseline？

great work! and Does it support multiple GPUs?