Ross Wightman
be7c784d21
Small TResNet simplification, just use SelectAdaptivePool, don't notice any perf difference
5 years ago
Ross Wightman
e3a98171b2
Remove some CRLF that made it into TResNet merge
5 years ago
Ross Wightman
0004f37d25
Some TResNet cleanup.
...
* allow use of global pool arg, test-time-pooling
* clean checkpoints to just contain state dict, add 448 res checkpoints
* support DataParallel via lazy filter creation for JIT Downsample
* some minor formatting (mostly alignment) preferences
5 years ago
Ross Wightman
64fe37d008
Remove old mean/std helper, rely fully on cmd line or default_cfg now. Fixes #126
5 years ago
Ross Wightman
e6f24e5578
Add 'momentum' optimizer (SGD w/o nesterov) for stable EfficientDet training defaults
5 years ago
Ross Wightman
ebf82b84ac
Merge pull request #122 from mrT23/master
...
TResNet models
5 years ago
Alexey Chernov
bdb165a8a4
Merge changes in feature extraction interface to MobileNetV3
...
Experimental feature extraction interface seems to be changed
a little bit with the most up to date version apparently found
in EfficientNet class. Here these changes are added to
MobileNetV3 class to make it support it and work again, too.
5 years ago
talrid
8a63c1add8
finalizing
5 years ago
talrid
6209146738
TResNet models
5 years ago
Ross Wightman
13cf68850b
Remove poorly named metrics from torch imagenet example origins. Use top1/top5 in csv output for consistency with existing validation results files, acc elsewhere. Fixes #111
5 years ago
Ross Wightman
56608c9070
Bump version for pypi
5 years ago
Ross Wightman
1a8f5900ab
Update EfficientNet feature extraction for EfficientDet. Add needed MaxPoolSame as well.
5 years ago
Chris Ha
06a50a94a8
Fix minor typos in create_attn.py and resnet.py
...
'eca'->'ceca'
and
doest not-> does not
5 years ago
Ross Wightman
c99a5abed4
Merge pull request #115 from rwightman/mobilenetv2-experiment
...
MobileNet-V2 experiments
5 years ago
Ross Wightman
e34074b4da
Add final weights for MobileNet-V2 experiments
5 years ago
Ross Wightman
a6a5565de7
Fiddling...
5 years ago
Ross Wightman
5a16c533ff
Add better resnext50_32x4d weights trained by andravin
5 years ago
Ross Wightman
bc998cad91
Experimenting with some MobileNetV2 variations to compare against EfficientNet-Lite
5 years ago
Ross Wightman
3406e582cf
Add EfficientNet-Lite results, update README
5 years ago
Ross Wightman
bd05258f7b
EfficientNet-Lite model added w/ converted checkpoints, validation in progress...
5 years ago
Andrew Lavin
d92cc4da54
Modified save_checkpoint to always save last checkpoint. Fixes #98 .
5 years ago
Ross Wightman
56e2ac3a6d
Merge pull request #94 from rwightman/lr_noise
...
Learning rate noise, MobileNetV3 weights, and activate MobileNetV3/EfficientNet weight init change
5 years ago
Ross Wightman
c16f25ced2
Add MobileNetV3 Large weights, results, update README and sotabench for merge
5 years ago
Ross Wightman
c60069c1eb
Annotate types on drop fns to avoid torchscript error
5 years ago
Ross Wightman
9fee316752
Enable fixed fanout calc in EfficientNet/MobileNetV3 weight init by default. Fix #84
5 years ago
Ross Wightman
27b3680d49
Revamp LR noise, move logic to scheduler base. Fixup PlateauLRScheduler and add it as an option.
5 years ago
Ross Wightman
cc5a11abba
version bump for PyPi update
5 years ago
Ross Wightman
514b0938c4
Experimenting with per-epoch learning rate noise w/ step scheduler
5 years ago
Ross Wightman
43225d110c
Unify drop connect vs drop path under 'drop path' name, switch all EfficientNet/MobilenetV3 refs to 'drop_path'. Update factory to handle new drop args.
5 years ago
Ross Wightman
f1d5f8a6c4
Update comments for Selective Kernel and DropBlock/Path impl, add skresnet34 weights
5 years ago
Ross Wightman
569419b38d
Tweak some comments, add SKNet models with weights to sotabench, remove an unused branch
5 years ago
Ross Wightman
53c47479c4
Batch validation batch size adjustment, tweak L2 crop pct
5 years ago
Ross Wightman
08553e16b3
Merge branch 'master' into attention
5 years ago
Ross Wightman
fa38f24967
Update SK network configs, add weights for skresnet8 and skresnext50
5 years ago
Ross Wightman
f098fda2ca
Add map_location='cpu' to ModelEma resume, should improve #72
5 years ago
Ross Wightman
ba15ca47e8
Add ported EfficientNet-L2, B0-B7 NoisyStudent weights from TF TPU
5 years ago
Ross Wightman
5e6dbbaf30
Add CBAM for experimentation
5 years ago
Ross Wightman
d725991870
Remove debug print from ECA module
5 years ago
Ross Wightman
2a7d256fd5
Re-enable mem-efficient/jit activations after torchscript tests
5 years ago
Ross Wightman
f902bcd54c
Layer refactoring continues, ResNet downsample rewrite for proper dilation in 3x3 and avg_pool cases
...
* select_conv2d -> create_conv2d
* added create_attn to create attention module from string/bool/module
* factor padding helpers into own file, use in both conv2d_same and avg_pool2d_same
* add some more test eca resnet variants
* minor tweaks, naming, comments, consistency
5 years ago
Ross Wightman
a99ec4e7d1
A bunch more layer reorg, splitting many layers into own files. Improve torchscript compatibility.
5 years ago
Ross Wightman
13746a33fc
Big move, layer modules and fn to timm/models/layers
5 years ago
Ross Wightman
f54612f648
Merge branch 'select_kernel' into attention
5 years ago
Ross Wightman
4defbbbaa8
Fix module name mistake, start layers sub-package
5 years ago
Ross Wightman
7011cd0902
A little bit of ECA cleanup
5 years ago
Ross Wightman
46471df7b2
Merge pull request #82 from VRandme/eca
...
ECA-Net Efficient Channel Attention
5 years ago
Ross Wightman
d0eb59ef46
Remove unused default_init for EfficientNets, experimenting with fanout calc for #84
5 years ago
Chris Ha
e6a762346a
Implement Adaptive Kernel selection
...
When channel size is given,
calculate adaptive kernel size according to original paper.
Otherwise use the given kernel size(k_size), which defaults to 3
5 years ago
Ross Wightman
13e8da2b46
SelectKernel split_input works best when input channels split like grouped conv, but output is full width. Disable zero_init for SK nets, seems a bad combo.
5 years ago
Chris Ha
6db087a1ff
Merge remote-tracking branch 'upstream/master' into eca
5 years ago