-
We can decrease memory requirements a lot by not performing the transposes necessary to convert from fortran to C order. Ran Lu for instance wants to work with Fortran order anyway so it's just wasted…
-
currently, the affinitymap mask needs to be precomputed and it is most IO cost. We should be able to compute it dynamically.
We have to focus the logic of different layers. make it right seems hard…
-
affinitymap is large and expensive to store, it would be good to continue running the watershed and drop the affinitymap.
-
I load cracker_60.pth checkpoint and try the below image, but have no cube detected. The result list is empty, and I have tried showing belief maps and affinity maps but they look fail. Do you have an…
-
the operators should correspond with image type.
-
## based on Funkey's waterz (easiest)
it already have python interface
need to test whether it work with our affinitymap or not
https://github.com/funkey/waterz
## based on latest C++ code (pref…
-
currently, the chunk will be masked out by affinitymap mask was still computed since the inference happens before the affinitymap mask download!
-
many functions should be able to be reused.
add another script to mask out affinity map.
abstract the interface to process a chunk, and ignoring the processing detail.
the interface should be input…
-
_Original author: mfschwa...@google.com (April 25, 2012 15:19:59)_
From messages posted by a gsutil user on gs-discussion and gsutil-discuss:
I have a set of files such that:
python gsutil ls gs://m…
-
affinitymap and semantic map is close to binary, but was formatted as Float32. In Phase II, this will cost about 2*4*(3+4) = 56 peterbyte without compression! We got about 50% compression rate using `…