Add inference script

[gachiemchiep]

Hello @HuAndrew
I add an inference script to run the trained model on video.

For the pose_hrnet_w32_student_FPD.pth trained model, I can inference fine using network definition at experiments/coco/hrnet/w32_256x192_adam_lr1e-3.yaml.

Unfortunately, the hourglass_student_FPD_.pth and hourglass_student_FPD.pth couldn't be loaded by experiments/fpd_mpii/hourglass/hg8_256x256_d256x3_adam_lr2.5e-4.yaml.
Where can I find the network definition for hourglass_student_FPD_.pth and hourglass_student_FPD.pth?

RuntimeError: Error(s) in loading state_dict for HourglassNet:
	size mismatch for conv1.weight: copying a param with shape torch.Size([32, 3, 7, 7]) from checkpoint, the shape in current model is torch.Size([64, 3, 7, 7]).
	size mismatch for conv1.bias: copying a param with shape torch.Size([32]) from checkpoint, the shape in current model is torch.Size([64]).
	size mismatch for bn1.weight: copying a param with shape torch.Size([32]) from checkpoint, the shape in current model is torch.Size([64]).
	size mismatch for bn1.bias: copying a param with shape torch.Size([32]) from checkpoint, the shape in current model is torch.Size([64]).
	size mismatch for bn1.running_mean: copying a param with shape torch.Size([32]) from checkpoint, the shape in current model is torch.Size([64]).