240e667746
Revert "mixer_b16_224_miil"
2021-05-20 10:23:07 +03:00
d046498e0b
update test_models.py
2021-05-20 11:20:39 +08:00
7b799c4e79
add latest code
2021-05-20 11:15:49 +08:00
cf4ce2f15b
Merge pull request #647 from rwightman/more_mlp
...
Add preliminary gMLP and ResMLP impl to Mlp-Mixer
2021-05-19 12:47:12 -07:00
b3b503c13d
Merge pull request #626 from cvcomar/master
...
fix: the exception not using default topk argument
2021-05-19 11:51:59 -07:00
6d81374b88
Update tests for new mlp models
2021-05-19 11:09:42 -07:00
5bcf686cb0
mixer_b16_224_miil
2021-05-19 20:51:10 +03:00
d5af752117
Add preliminary gMLP and ResMLP impl to Mlp-Mixer
2021-05-19 09:55:05 -07:00
00548b8427
Add Twins
2021-05-18 19:21:53 +08:00
e7f0db8664
Fix drop/drop_path arg on MLP-Mixer model. Fix #641
2021-05-16 08:31:52 -07:00
9a3ae97311
Another set of byoanet models w/ ECA channel + SA + groups
2021-05-15 22:56:51 -07:00
d53e91218e
Fix tf.data options setting for newer TF versions
2021-05-15 22:56:12 -07:00
7077f16c6a
Change 21k model naming from _21k to _in21k for consistency with existing 21k models.
2021-05-15 12:42:26 -07:00
94d4b53352
Add temporary default_cfgs to visformer models so they pass tests
2021-05-15 08:41:31 -07:00
2a72d38ba2
Update README.md
2021-05-14 23:06:49 -07:00
3bffc701f1
Merge branch 'master' into levit_visformer_rednet
2021-05-14 23:02:12 -07:00
758802e1f6
Merge pull request #636 from rwightman/effnetv2_official
...
EfficientNetV2 official impl w/ weights ported from TF.
2021-05-14 17:24:11 -07:00
ecc7552c5c
Add levit, levit_c, and visformer model defs. Largely untested and not finished cleanup.
2021-05-14 17:16:34 -07:00
165fb354b2
Add initial RedNet model / Involution layer impl for testing
2021-05-14 17:16:34 -07:00
328249f11a
Update README, tweak fine-tune effv2 model names.
2021-05-14 16:37:43 -07:00
c4f482a08b
EfficientNetV2 official impl w/ weights ported from TF. Cleanup/refactor of related EfficientNet classes and models.
2021-05-14 15:50:00 -07:00
c16d65a8a7
Update results.csv files with cait, mlp-mixer, resnetrs, miil imagenet21k pretrained, and my latest efficientnet v2s/b4 weights.
2021-05-11 10:50:46 -07:00
4fbc32d3d0
Fix crop_pct for cait models.
2021-05-11 10:50:15 -07:00
d7e1e7144a
fix: the exeption when topk is 1
2021-05-10 16:52:40 +09:00
ffa90e04d3
fix: the exception not using default topk argument
2021-05-10 16:20:19 +09:00
715519a5ef
Rethink name of patch embed grid info
2021-05-06 14:08:20 -07:00
b2c305c2aa
Move Mlp and PatchEmbed modules into layers. Being used in lots of models now...
2021-05-06 14:03:23 -07:00
3ba6b55cb2
More adjustments to ByoaNet models for further experiments.
2021-05-06 13:59:44 -07:00
f480734ad5
Update README.md
2021-05-05 23:17:35 -07:00
504fef4bc1
Merge pull request #609 from rwightman/cait
...
CaiT transformer model
2021-05-05 18:00:22 -07:00
d45e50b9db
Update test for cait 448x448 model
2021-05-05 17:51:23 -07:00
5fcddb96a8
Merge branch 'master' into cait
2021-05-05 17:29:38 -07:00
3db12b4b6a
Finish CaiT cleanup
2021-05-05 17:28:19 -07:00
936e9b331d
Merge pull request #613 from rwightman/mlp-mixer
...
MLP-Mixer models w/ pretrained weights ported from Google JAX impl...
2021-05-05 16:59:29 -07:00
2d8b09fe8b
Add official pretrained weights to MLP-Mixer, complete model cfgs.
2021-05-05 15:59:40 -07:00
12efffa6b1
Initial MLP-Mixer attempt...
2021-05-05 00:59:45 -07:00
0721559511
Improved (hopefully) init for SA/SA-like layers used in ByoaNets
2021-05-04 21:40:39 -07:00
d5473c17f7
Fix incorrect name of shortcut/identity paths in many residual nets. Inherited from naming in old old torchvision, long fixed there.
2021-05-04 21:27:15 -07:00
0d87650fea
Remove filter hack from BlurPool w/ non-persistent buffer. Use BlurPool2d instead of AntiAliasing.. for TResNet. Breaks PyTorch < 1.6.
2021-05-04 16:56:28 -07:00
ddc743fdf8
Update ResNet-RS models to EMA weights
2021-05-04 16:16:55 -07:00
08d60f4a9a
resnetrs50 pool sizing wrong
2021-05-04 12:41:09 -07:00
137a374930
Merge pull request #555 from MichaelMonashev/patch-1
...
benchmark.py argument description fixed
2021-05-04 11:44:01 -07:00
1daa15ecc3
Initial Cait commit. Still some cleanup to do.
2021-05-04 11:19:27 -07:00
67d0665b46
Post ResNet-RS merge cleanup. Add weight urls, adjust train/test/crop pct.
2021-05-04 11:04:23 -07:00
560eae38f5
[WIP] Add ResNet-RS models ( #554 )
...
* Add ResNet-RS models
* Only include resnet-rs changes
* remove whitespace diff
* EOF newline
* Update time
* increase time
* Add first conv
* Try running only resnetv2_101x1_bitm on Linux runner
* Add to exclude filter
* Run test_model_forward_features for all
* Add to exclude ftrs
* back to defaults
* only run test_forward_features
* run all tests
* Run all tests
* Add bigger resnetrs to model filters to fix Github CLI
* Remove resnetv2_101x1_bitm from exclude feat features
* Remove hardcoded values
* Make sure reduction ratio in resnetrs is 0.25
* There is no bias in replaced maxpool so remove it
2021-05-04 10:59:44 -07:00
9cc7dda6e5
Fixup byoanet configs to pass unit tests. Add swin_attn and swinnet26t model for testing.
2021-04-29 21:08:37 -07:00
e15c3886ba
Defaul lambda r=7. Define '26t' stage 4/5 256x256 variants for all of bot/halo/lambda nets for experiment. Add resnet50t for exp. Fix a few comments.
2021-04-29 10:58:49 -07:00
d15ad3e919
Merge branch 'morizin-master'
2021-04-28 18:10:12 -07:00
e5e15754c9
Fix coat first conv ident
2021-04-28 18:09:23 -07:00
76739a7589
CoaT merge. Bit of formatting, fix torchscript (for non features), remove einops/einsum dep, add pretrained weight hub (url) support.
2021-04-28 16:31:35 -07:00
talrid
李鑫杰
Ross Wightman
Ross Wightman
comar
Aman Arora