astuti / ila-da Goto Github PK

I have read your paper Instance Level Affinity Based Transfer for Unsupervised Domain Adaptation, and I am very interested in this paper.
I hope you can provide the code as soon as possible, which will be very helpful to me. Thank you very much

Hi, thanks for your good work. And there seems to be something wrong in the criterion_factory.py.
① One variable is missing: What is the self.n_per_domain in Line 22, 26, 48? And what does this mean?
② The def (__target_pseudolabels) doesn't seem to work. When I try to give the value 10 to the self.n_per_domain, an error occurs in Line 27. shape mismatch: indexing tensors could not be broadcast together with shapes [80], [10]. (bs is set to 80).
Where sim_matrix is the shape [80,80],
ind is the shape [80,80],
ind_split is a list with the length 80 (the i'th row represents the top-i target samples compared to all source samples.),
vr_src: [80, 10]
_row: [80]
_col: [10]

Thanks for your reply.

Hi, nice work !
Can you share some links to the dataset ?
Thank you very much !