mm
0 Followers · 0 Following
  • Joined on 2022-01-17
Repositories
12
 Projects Packages Public Activity Starred Repositories
mm pushed to main at mm/colors 2024-02-14 16:10:03 +00:00
a2331e0b75 more spacing
mm pushed to main at mm/colors 2024-02-14 16:06:05 +00:00
f0a4c940af bg filling
04ff370bbb wip
90f08a5edb paths
831f17a9b4 tweaks + path changes
698f6f1b51 seeds
Compare 81 commits »
mm pushed to new-sep-loss at mm/colors 2024-02-14 16:05:48 +00:00
f0a4c940af bg filling
mm pushed to new-sep-loss at mm/colors 2024-02-14 04:55:47 +00:00
04ff370bbb wip
90f08a5edb paths
Compare 2 commits »
mm pushed to new-sep-loss at mm/colors 2024-02-13 06:55:24 +00:00
831f17a9b4 tweaks + path changes
mm pushed to new-sep-loss at mm/colors 2024-02-13 06:46:54 +00:00
698f6f1b51 seeds
mm pushed to new-sep-loss at mm/colors 2024-02-13 06:46:24 +00:00
13dd754ba6 nice set of params
a97055f282 GNU parallel
d1b963e05d minimal requirements.
Compare 3 commits »
mm pushed to new-sep-loss at mm/colors 2024-02-13 05:28:34 +00:00
f3bc1e90d3 ignore npy
b95cb8df3d cuda libs
527cc285f3 drop in cuml
Compare 3 commits »
mm pushed to new-sep-loss at mm/colors 2024-02-13 03:50:46 +00:00
95a6adac91 lint
mm pushed to new-sep-loss at mm/colors 2024-02-13 03:42:57 +00:00
447b773d03 no annotate, update target
00cba285c8 minor tweaks, check annotations
9953327d30 copy over color files, use new plotting
Compare 3 commits »
mm pushed to new-sep-loss at mm/colors 2024-01-28 09:54:40 +00:00
0ae8414481 syntax, no diff
mm pushed to new-sep-loss at mm/colors 2024-01-28 09:43:40 +00:00
2af491c324 tweak metric
mm pushed to new-sep-loss at mm/colors 2024-01-28 08:32:14 +00:00
5687f30818 back to "correct" metric
5def982f12 recreate iris looking result. weird though.
Compare 2 commits »
mm pushed to new-sep-loss at mm/colors 2024-01-28 07:31:56 +00:00
a9e772f34e restrict experiment, add loop option + dropout
mm pushed to new-sep-loss at mm/colors 2024-01-28 06:34:31 +00:00
697030b7df partial unsupervision
mm pushed to new-sep-loss at mm/colors 2024-01-28 06:33:26 +00:00
f450611ce3 pin requirements
mm pushed to new-sep-loss at mm/colors 2024-01-28 06:05:35 +00:00
0c4990e98c remove depth = 8 and switch to relu
mm pushed to new-sep-loss at mm/colors 2024-01-28 05:26:23 +00:00
05dd4e29ce sequential instead of loopback
mm pushed to new-sep-loss at mm/colors 2024-01-28 02:58:56 +00:00
d318480b7c circle norm bad for supervise?
1c116f3f12 benchmark supervised again
c7ffd09fb4 lr scheduler
Compare 3 commits »
mm pushed to new-sep-loss at mm/colors 2024-01-28 02:43:39 +00:00
4342a54cc8 anchors and more train data, exp w batch
248d1a72f9 re-anchor, pretty meh results in-batch
9e4861a272 try unsupervised again, but with 10k random samples
e5b6f287a3 xkcd colors may be too few to learn from. need 10x
865e7f5104 supervised questionable
Compare 5 commits »