search

princeton-vl / pose-ae-train

Training code for "Associative Embedding: End-to-End Learning for Joint Detection and Grouping"
BSD 3-Clause "New" or "Revised" License
373 stars 76 forks source link

Question about the result of the forwardNet #46

Closed 1994cxy closed 4 years ago

1994cxy commented 5 years ago

Thanks for the excellent work, but I have some questions about the result of the forwardNet. I think the result of the forwardNet should be batch_size channel output_res output_res, and we should use result[:,:17,:,:] for detection loss and result[:,17:34,:,:] for embedding loss. However, in models/posenet.py line 52, the author uses dets = preds[:,:,:17] and tags = preds[:,:,17:34] It seems like the result of forwardNet is output_resoutput_res*channel. Way? Does this has anything to do with line 49 in posenet.py?

1994cxy commented 5 years ago

@anewell

1994cxy commented 4 years ago

I think I may find the answer, preds[:,:,:17] is equal to preds[:,:,:17,:]. Sorry for bothering.