Ross Wightman
|
4d7a5544f7
|
Remove inplace sigmoid for consistency with other impl
|
3 years ago |
|
Ross Wightman
|
88a5b54802
|
A few small evonorm tweaks for convergence comparisons
|
3 years ago |
|
Ross Wightman
|
66daee4f31
|
Last change wasn't complete, missed adding full evo_norm changeset
|
3 years ago |
|
Ross Wightman
|
7bbbd5ef1b
|
EvoNorm and GroupNormAct options for debugging TPU / XLA concerns
|
3 years ago |
|
Ross Wightman
|
ff0f709c20
|
Testing TFDS shuffle across epochs
|
3 years ago |
|
Ross Wightman
|
69e90dcd8c
|
Merge branch 'norm_norm_norm' into bits_and_tpu
|
3 years ago |
|
Ross Wightman
|
d04f2f1377
|
Update drop_path and drop_block (fast impl) to be symbolically traceable, slightly faster
|
3 years ago |
|
Ross Wightman
|
820ae9925e
|
Fix load_state_dict to handle None ema entries
|
3 years ago |
|
Ross Wightman
|
0e212e8fe5
|
Merge remote-tracking branch 'origin/norm_norm_norm' into bits_and_tpu
|
3 years ago |
|
Ross Wightman
|
cd059cbe9c
|
Add FX backward tests back
|
3 years ago |
|
Ross Wightman
|
834a9ec721
|
Disable use of timm nn.Linear wrapper since AMP autocast + torchscript use appears fixed
|
3 years ago |
|
Ross Wightman
|
cad170e494
|
Merge remote-tracking branch 'origin/norm_norm_norm' into bits_and_tpu
|
3 years ago |
|
Ross Wightman
|
58ffa2bfb7
|
Update pytest for GitHub runner to use --forked with xdist, hopefully eliminate memory buildup
|
3 years ago |
|
Ross Wightman
|
78912b6375
|
Updated EvoNorm implementations with some experimentation. Add FilterResponseNorm. Updated RegnetZ and ResNetV2 model defs for trials.
|
3 years ago |
|
Ross Wightman
|
55adfbeb8d
|
Add commented code to increase open file limit via Python (for TFDS dataset building)
|
3 years ago |
|
Ross Wightman
|
f7d210d759
|
Remove evonorm models from FX tests
|
3 years ago |
Ross Wightman
|
3dc71695bf
|
Merge pull request #989 from martinsbruveris/feat/resmlp-dino
Added DINO pretrained ResMLP models.
|
3 years ago |
|
Ross Wightman
|
480c676ffa
|
Fix FX breaking assert in evonorm
|
3 years ago |
Martins Bruveris
|
85c5ff26d7
|
Added DINO pretrained ResMLP models.
|
3 years ago |
|
Ross Wightman
|
f83b0b01e3
|
Would like to pass GitHub tests again disabling both FX feature extract backward and torchscript tests
|
3 years ago |
|
Ross Wightman
|
a22b85c1b9
|
Merge branch 'nateraw-hf-save-and-push'
|
3 years ago |
|
Ross Wightman
|
d633a014e6
|
Post merge cleanup. Fix potential security issue passing kwargs directly through to serialized web data.
|
3 years ago |
|
Ross Wightman
|
8a83c41d7b
|
Merge branch 'hf-save-and-push' of https://github.com/nateraw/pytorch-image-models into nateraw-hf-save-and-push
|
3 years ago |
|
Ross Wightman
|
147e1059a8
|
Remove FX backward test from GitHub actions runs for now.
|
3 years ago |
Nathan Raw
|
b18c9e323b
|
Update helpers.py
|
3 years ago |
|
Nathan Raw
|
308d0b9554
|
Merge branch 'master' into hf-save-and-push
|
3 years ago |
|
Ross Wightman
|
878bee1d5e
|
Add patch8 vit model to FX exclusion filter
|
3 years ago |
|
Ross Wightman
|
9bb4c80d2a
|
Update README.md, missed recent regnetz results
|
3 years ago |
|
Ross Wightman
|
79bf4f163f
|
Add mention of ResNet50 weights w/ RSB recipe ingredients and lower train res for better @ 224 results (but worse res scaling beyond). Not changing default.
|
3 years ago |
|
Ross Wightman
|
ce76a810c2
|
New FX test strategy, filter based on param count
|
3 years ago |
|
Ross Wightman
|
1e51c2d02e
|
More FX test tweaks
|
3 years ago |
|
Ross Wightman
|
f0507f6da6
|
Fix k_decay default arg != 1.0 in poly scheduler
|
3 years ago |
|
Ross Wightman
|
947e1df3ef
|
Update README.md
|
3 years ago |
|
Ross Wightman
|
90448031ea
|
Filter more large models from FX tests
|
3 years ago |
|
Ross Wightman
|
1f53db2ece
|
Updated lamhalobotnet weights, 81.5 top-1
|
3 years ago |
|
Ross Wightman
|
8dc269c303
|
Filter more models for FX tests
|
3 years ago |
|
Ross Wightman
|
15ef108eb4
|
Add better halo2botnet50ts weights, 82 top-1 @ 256
|
3 years ago |
|
Ross Wightman
|
2482652027
|
Add nfnet_f2 to FX test exclusion
|
3 years ago |
|
Ross Wightman
|
809c7bb1ec
|
Merge remote-tracking branch 'origin/master' into bits_and_tpu
|
3 years ago |
|
Ross Wightman
|
734b2244fe
|
Add RegNetZ-D8 (83.5 @ 256, 84 @ 320) and RegNetZ-E8 (84.5 @ 256, 85 @ 320) weights. Update names of existing RegZ models to include group size.
|
3 years ago |
|
Ross Wightman
|
05092e2fbe
|
Add more models to FX filter
|
3 years ago |
|
Ross Wightman
|
93cc08fdc5
|
Make evonorm variables 1d to match other PyTorch norm layers, will break weight compat for any existing use (likely minimal, easy to fix).
|
3 years ago |
|
Ross Wightman
|
af607b75cc
|
Prep a set of ResNetV2 models with GroupNorm, EvoNormB0, EvoNormS0 for BN free model experiments on TPU and IPU
|
3 years ago |
|
Ross Wightman
|
3819bef93e
|
Add FX test exclusion since it uses more ram and barfs on GitHub actions. Will take a few iterations to include needed models :(
|
3 years ago |
|
Ross Wightman
|
c976a410d9
|
Add ResNet-50 w/ GN (resnet50_gn) and SEBotNet-33-TS (sebotnet33ts_256) model defs and weights. Update halonet50ts weights w/ slightly better variant in1k val, more robust to test sets.
|
3 years ago |
|
Ross Wightman
|
9b3519545d
|
Attempt to reduce memory footprint of FX tests for GitHub actions runs
|
3 years ago |
|
Ross Wightman
|
bdd3dff0ca
|
beit_large models killing GitHub actions test, filter out
|
3 years ago |
|
Ross Wightman
|
9d6aad44f8
|
Update tests to run Python 3.9, PyTorch 1.10, torchvision 0.11.1
|
3 years ago |
|
Ross Wightman
|
ee40b582bb
|
Merge branch 'alexander-soare-fx-feature-extract-new'
|
3 years ago |
|
Ross Wightman
|
f2006b2437
|
Cleanup qkv_bias cat in beit model so it can be traced
|
3 years ago |
