Commit graph

Select branches

Hide pull requests

main

#1

Mono Color

• f080af0b57
  
  identity test: plot binarised heightmap Starbeamrainbowlabs 2023-01-06 19:03:06 +0000
• 36859746ff
  
  dlr: fix crash Starbeamrainbowlabs 2023-01-06 17:13:35 +0000
• bcf198d47b
  
  mono identity test: output Starbeamrainbowlabs 2023-01-06 17:08:18 +0000
• db0b010814
  
  slur dlr: log file names correct Starbeamrainbowlabs 2023-01-05 19:47:51 +0000
• e01ecfb615
  
  slurm dlr: fix output dir Starbeamrainbowlabs 2023-01-05 19:42:42 +0000
• aa76d754c1
  
  slurm dlr: fix pathing Starbeamrainbowlabs 2023-01-05 19:35:56 +0000
• 67b8a2c6c0
  
  spaces → spaces Starbeamrainbowlabs 2023-01-05 19:17:44 +0000
• 56a501f8a9
  
  weights="imagenet" only works with 3 image channels Starbeamrainbowlabs 2023-01-05 19:09:31 +0000
• 4563fe6b27
  
  dpl: fix moar crashes Starbeamrainbowlabs 2023-01-05 19:03:44 +0000
• fefeb5d531
  
  fix water depth fiddling Starbeamrainbowlabs 2023-01-05 19:01:20 +0000
• 46d1f5e4e0
  
  dataset_mono: fix dataset parsing Starbeamrainbowlabs 2023-01-05 19:00:52 +0000
• aca7b83a78
  
  dataset_mono: fix sizing it didn't account for rainfall_scale_up Starbeamrainbowlabs 2023-01-05 18:38:47 +0000
• 19bb2fcac0
  
  debug Starbeamrainbowlabs 2023-01-05 18:32:22 +0000
• 6a4f68a055
  
  missing import Starbeamrainbowlabs 2023-01-05 18:26:33 +0000
• 0d4cc63b76
  
  dl rainfall: fix env var name Starbeamrainbowlabs 2023-01-05 17:42:20 +0000
• 9d87168edb
  
  Merge branch 'main' of git.starbeamrainbowlabs.com:sbrl/PhD-Rainfall-Radar Starbeamrainbowlabs 2023-01-05 17:11:01 +0000
• dd79fb6e68
  
  fixup Starbeamrainbowlabs 2023-01-05 17:09:09 +0000
• 11ccd4cbee
  
  slurm deeplab rainfall: fix variable naming Starbeamrainbowlabs 2023-01-05 17:08:57 +0000
• 3df5cdb45d
  
  Slurm: finish writing for DeepLabV3+ rainfall Starbeamrainbowlabs 2022-12-16 20:21:42 +0000
• c17e53ca75
  
  deeplabv3+ for rainfall Starbeamrainbowlabs 2022-12-16 19:52:59 +0000
• 677e39f820
  
  work on slurm for deeplabv3+ rainfall, but it's NOT FINISHED YET Starbeamrainbowlabs 2022-12-16 19:52:44 +0000
• 423e277ed1
  
  add comment Starbeamrainbowlabs 2022-12-15 19:33:25 +0000
• ef5071b569
  
  DeepLabV3+: start working on version for rainfall radar, but it's not finished yet Starbeamrainbowlabs 2022-12-15 19:33:14 +0000
• 15a3150127
  
  DeepLabV3+: close each matplotlib figure after writing it Starbeamrainbowlabs 2022-12-15 19:14:07 +0000
• 6ce121f861
  
  DeepLabV3+: have argument for number of channels Starbeamrainbowlabs 2022-12-14 17:36:30 +0000
• 1dc2ec3a46
  
  DeepLabV3+: pathing.... again Starbeamrainbowlabs 2022-12-13 18:51:09 +0000
• eb47f8f544
  
  dataset_mono: adjust to suit DeepLabV3+ too Starbeamrainbowlabs 2022-12-13 18:37:38 +0000
• 440e693dfc
  
  DeepLabv3+: fix pathing again Starbeamrainbowlabs 2022-12-13 18:26:00 +0000
• 7e1f271bf4
  
  deeplabv3+: fix colourmap Starbeamrainbowlabs 2022-12-13 14:02:10 +0000
• 4d8ce792c9
  
  ddeeplabv3+: fix imports/pathing errors Starbeamrainbowlabs 2022-12-13 13:38:27 +0000
• fc43f145c2
  
  if not Starbeamrainbowlabs 2022-12-13 13:28:09 +0000
• d907dc48e5
  
  DeepLabv3+: add logging Starbeamrainbowlabs 2022-12-13 13:20:16 +0000
• be4d928319
  
  deeplabv3+: chmod +x Starbeamrainbowlabs 2022-12-13 13:06:42 +0000
• 91846079b2
  
  deeplabv3+ tesst: add shebang Starbeamrainbowlabs 2022-12-13 12:56:14 +0000
• 96e260fe82
  
  slurm: add job file for deeplabv3 test Starbeamrainbowlabs 2022-12-12 19:31:49 +0000
• 8866960017
  
  TEST SCRIPT: deeplabv3 ref https://keras.io/examples/vision/deeplabv3_plus/ dataset ref https://drive.google.com/uc?id=1B9A9UCJYMwTL4oBEo4RZfbMZMaZhKJaz Starbeamrainbowlabs 2022-12-12 19:19:24 +0000
• 4e4d42a281
  
  LossDice: add comment Starbeamrainbowlabs 2022-12-12 18:34:20 +0000
• 449bc425a7
  
  LossDice: explicitly cast inputs to float32 Starbeamrainbowlabs 2022-12-12 17:20:32 +0000
• dbf8f5617c
  
  drop activation function in last layers Starbeamrainbowlabs 2022-12-12 17:20:04 +0000
• bcd2f1251e
  
  LossDice: Do 1 - thing instead of -thing Starbeamrainbowlabs 2022-12-09 19:41:32 +0000
• d0dbc50bb7
  
  debug Starbeamrainbowlabs 2022-12-09 19:33:28 +0000
• 2142bb039c
  
  again Starbeamrainbowlabs 2022-12-09 19:30:01 +0000
• 7000b0f193
  
  fixup Starbeamrainbowlabs 2022-12-09 19:23:35 +0000
• 85012d0616
  
  fixup Starbeamrainbowlabs 2022-12-09 19:18:03 +0000
• 719d8e9819
  
  strip channels layer at end Starbeamrainbowlabs 2022-12-09 19:11:00 +0000
• 0129c35a35
  
  LossDice: remove weird K.* functions Starbeamrainbowlabs 2022-12-09 19:06:26 +0000
• 659fc97fd4
  
  fix crash Starbeamrainbowlabs 2022-12-09 18:39:27 +0000
• e22c0981e6
  
  actually use dice loss Starbeamrainbowlabs 2022-12-09 18:35:17 +0000
• 649c262960
  
  mono: switch loss from crossentropy to dice Starbeamrainbowlabs 2022-12-09 18:13:37 +0000
• 7fd7c750d6
  
  jupyter: identity test status: FAILED, as usual....! Don't worry though, 'cause we has a *planses*..... MUHAHAHAHAHAHAHA * cue evil laugh * Starbeamrainbowlabs 2022-12-09 18:07:56 +0000
• cf9e8aa237
  
  jupyter: convnext-mono identity test Starbeamrainbowlabs 2022-12-09 15:50:27 +0000
• 2a1772a211
  
  confvnext_intrevse: add shallow Starbeamrainbowlabs 2022-12-08 19:10:12 +0000
• c27869630a
  
  I hate VSCode's git commit interface it doesn't let you ammend Starbeamrainbowlabs 2022-12-08 18:58:54 +0000
• b3345963f3
  
  missing arg pass Starbeamrainbowlabs 2022-12-08 18:58:32 +0000
• 3dde9b69da
  
  fixup Starbeamrainbowlabs 2022-12-08 18:56:32 +0000
• 6fce39f696
  
  WHY?!?!?! Starbeamrainbowlabs 2022-12-08 18:55:53 +0000
• 26766366fc
  
  I hate the python code intelligence it's bad Starbeamrainbowlabs 2022-12-08 18:55:15 +0000
• ff56f591c7
  
  I hate python Starbeamrainbowlabs 2022-12-08 18:53:37 +0000
• d37e7224f5
  
  train-mono: tidy up arg passing Starbeamrainbowlabs 2022-12-08 18:47:03 +0000
• b53db648bf
  
  fixup Starbeamrainbowlabs 2022-12-08 18:31:42 +0000
• 18c0210704
  
  typo Starbeamrainbowlabs 2022-12-08 17:00:25 +0000
• a3c9416cf0
  
  LossCrossentropy: don't sum Starbeamrainbowlabs 2022-12-08 16:57:11 +0000
• 08046340f4
  
  dataset_mono: normalise heightmap Starbeamrainbowlabs 2022-12-08 16:10:34 +0000
• d997157f55
  
  dataset_mono: log when using heightmap Starbeamrainbowlabs 2022-12-06 19:30:11 +0000
• 468c150570
  
  slurm-train-mono: add HEIGHTMAP Starbeamrainbowlabs 2022-12-06 19:28:06 +0000
• d0f2e3d730
  
  readfile: do transparent gzip by default ....but there's a glad to turn it off if needed Starbeamrainbowlabs 2022-12-06 19:27:39 +0000
• eac6472c97
  
  Implement support for (optionally) taking a heightmap in Starbeamrainbowlabs 2022-12-06 18:55:58 +0000
• f92b2b3472
  
  according to the equation it looks like it's 2 Starbeamrainbowlabs 2022-12-02 17:22:46 +0000
• cad82cd1bc
  
  CBAM: unsure if it's 1 ro 3 dense ayers in the shared mlp Starbeamrainbowlabs 2022-12-02 17:21:13 +0000
• 62f6a993bb
  
  implement CBAM, but it's UNTESTED Convolutional Block Attention Module. Starbeamrainbowlabs 2022-12-02 17:17:45 +0000
• 9d666c3b38
  
  train mono: type=int → float Starbeamrainbowlabs 2022-12-01 15:39:44 +0000
• 53dfa32685
  
  model_mono: log learning rate Starbeamrainbowlabs 2022-12-01 15:10:51 +0000
• c384d55dff
  
  add arg to adjust learning rate Starbeamrainbowlabs 2022-11-29 20:55:00 +0000
• 8e23e9d341
  
  model_segmenter: we're no longer using sparse Starbeamrainbowlabs 2022-11-29 19:28:27 +0000
• 9a2b4c6838
  
  dsseg: fix reshape/onehot ordering Starbeamrainbowlabs 2022-11-29 19:28:13 +0000
• df774146d9
  
  dataset_segmenter: reshape, not squeeze Starbeamrainbowlabs 2022-11-29 19:24:54 +0000
• 77b8a1a8db
  
  dataset_segmenter: squeeze Starbeamrainbowlabs 2022-11-29 19:16:15 +0000
• 2258b5a229
  
  slurm-train: reduce RAM required by 10GB Starbeamrainbowlabs 2022-11-29 19:15:34 +0000
• 01101ad30b
  
  losscrossentropy: return the reduced value * facepalm * Starbeamrainbowlabs 2022-11-29 19:07:08 +0000
• ff65393e78
  
  log file naming update Starbeamrainbowlabs 2022-11-29 18:41:14 +0000
• 37f196a785
  
  LossCrossentropy: add kwargs Starbeamrainbowlabs 2022-11-29 15:40:35 +0000
• 838ff56a3b
  
  mono: fix loading checkpoint Starbeamrainbowlabs 2022-11-29 15:25:11 +0000
• dba6cbffcd
  
  WHY. * facepalms * Starbeamrainbowlabs 2022-11-28 19:33:42 +0000
• 57b8eb93fb
  
  fixup Starbeamrainbowlabs 2022-11-28 19:09:35 +0000
• 6640a41bb7
  
  almost got it....? it's not what I expected....! Starbeamrainbowlabs 2022-11-28 19:08:50 +0000
• f48473b703
  
  fixup Starbeamrainbowlabs 2022-11-28 19:00:11 +0000
• f6feb125e3
  
  this iss ome serious debugging. This commit will produce an extremely large volume of output. Starbeamrainbowlabs 2022-11-28 18:57:41 +0000
• 09f81b0746
  
  train_mono: debug this commit will generate a large amount of debug output. Starbeamrainbowlabs 2022-11-28 16:46:17 +0000
• f39e4ade70
  
  LayerConvNextGamma: fix config serialisation bug .....this is unlikely to be the problem as this bug is in an unused code path. Starbeamrainbowlabs 2022-11-25 21:16:31 +0000
• e7410fb480
  
  train_mono_predict: limit label size to 64x64 that's the size the model predicts Starbeamrainbowlabs 2022-11-25 17:47:17 +0000
• 51dd484d13
  
  fixup Starbeamrainbowlabs 2022-11-25 16:55:45 +0000
• 884c4eb150
  
  rainfall_stats: formatting again Starbeamrainbowlabs 2022-11-24 19:08:07 +0000
• bfe038086c
  
  rainfall_stats: formatting Starbeamrainbowlabs 2022-11-24 19:07:44 +0000
• 7dba03200f
  
  fixup Starbeamrainbowlabs 2022-11-24 19:06:48 +0000
• e5258b9c66
  
  typo Starbeamrainbowlabs 2022-11-24 19:06:13 +0000
• 64d646bb13
  
  rainfall_stats: formatting Starbeamrainbowlabs 2022-11-24 19:05:35 +0000
• 675c7a7448
  
  fixup Starbeamrainbowlabs 2022-11-24 19:03:28 +0000
• afc1cdcf02
  
  fixup Starbeamrainbowlabs 2022-11-24 19:02:58 +0000
• e4bea89c89
  
  typo Starbeamrainbowlabs 2022-11-24 19:01:52 +0000
• a40cbe8705
  
  rainfall_stats: remove unused imports Starbeamrainbowlabs 2022-11-24 19:01:18 +0000