ppogg/retinaface_ghost

This is a project based on retinaface face detection, including ghostnet and mobilenetv3

retinaface_ghost's Issues

pre-trained model for mobilenetv3

Can I get pre-trained model for mobilenet v3 that is modified in this Retinaface_Ghost?
If so, I think that I can train my data efficiently.

Using ghostnet_Final.pth

when I use ghostnet_Final.pth, after running "python /content/Retinaface_Ghost/widerface_evaluate/evaluation.py" command, the result for all forms (easy, medium, hard) is zero for me, Can you direct me?

hey，I want to know what the pred_dir(in evaluation.py) folder's usage is and where can i find these files,thanks

Can I train two classes?

I want to add one more class in this network. How can I modify it?

pre-trained model for ghostnet6.7m

raise StopIteration StopIteration

Traceback (most recent call last): File "train.py", line 178, in train() File "train.py", line 134, in train images, targets = next(batch_iterator) File "/home/wyw/.conda/envs/py36_torch/lib/python3.6/site-packages/torch/utils/data/dataloader.py", line 435, in next data = self._next_data() File "/home/wyw/.conda/envs/py36_torch/lib/python3.6/site-packages/torch/utils/data/dataloader.py", line 1058, in _next_data raise StopIteration StopIteration

batch_iterator = iter(data.DataLoader(dataset, batch_size, shuffle=True, num_workers=num_workers,
collate_fn=detection_collate, drop_last=True))
change to
batch_iterator = iter(data.DataLoader(dataset, batch_size, shuffle=True, num_workers=num_workers,
collate_fn=detection_collate))

I train the resnet50, but test no result output, any other problem?

Want to know where the labels.txt files used in the author's training are generated?

Want to know where the labels.txt files used in the author's training are generated?

Onnx model does not correctly works

I converted the "mobilenet0.25_Final" model to onnx, but as a result of inference I get completely different results for the same image. Could you share your scripts with this onnx model?