jiaxuan liu
Results
3
issues of
jiaxuan liu
What's the motivation of using a one-layer-encoder instead of a deeper one?
add distributed data parallel in pytorch: add `with dev ` in correlation.py use cupy.RawModule instead of cupy.cuda.compile_with_cache for deprecation in cupy v10 add test_ddp for test
After the intrinsic calibration process, getting the intrinsic $K$ and destortion parameters $\sigma$, Should i use the undestorted images $\hat{I}$ and camera intrinsics $K$ as the ILCC input， or use...