Skip to content

Update timm universal (support transformer-style model) #1004

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 14 commits into from
Dec 19, 2024
Merged

Update timm universal (support transformer-style model) #1004

Changes from 1 commit
Commits
Show all changes
14 commits
Select commit Hold shift + click to select a range
363a361
Update timm_universal.py
brianhou0208 Dec 7, 2024
f07e107
Fix ruff style and typing
brianhou0208 Dec 7, 2024
ea725dd
Update encoders_timm.rst
brianhou0208 Dec 7, 2024
029c190
Fix typo error
brianhou0208 Dec 7, 2024
c58b7cf
Fix typo error & Update doc
brianhou0208 Dec 8, 2024
eae7e2b
Fix typo error
brianhou0208 Dec 8, 2024
4148788
Support channel-last format
brianhou0208 Dec 10, 2024
51a4d7b
Update encoders_timm.rst
brianhou0208 Dec 18, 2024
8b0fece
Update timm_universal.py
brianhou0208 Dec 18, 2024
330e6e5
Fix ruff style
brianhou0208 Dec 18, 2024
d8ea35f
Update timm_universal.py
brianhou0208 Dec 18, 2024
e7bc6e0
Add tests/test_models & fix type
brianhou0208 Dec 19, 2024
dd25aa2
Update test_models.py
brianhou0208 Dec 19, 2024
f55eb13
Update test_models.py
brianhou0208 Dec 19, 2024
File filter

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
Prev Previous commit
Next Next commit
Update encoders_timm.rst
  • Loading branch information
@brianhou0208
brianhou0208 committed Dec 18, 2024
commit 51a4d7bab379c02717e0da16a8a5ba553e349039
98 changes: 69 additions & 29 deletions docs/encoders_timm.rst
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -9,7 +9,7 @@ however, not all models are supported

Below is a table of suitable encoders (for DeepLabV3, DeepLabV3+, and PAN dilation support is needed also)

Total number of encoders: 792 (579+213)
Total number of encoders: 812 (593+219)

.. note::

Expand Down Expand Up @@ -99,6 +99,8 @@ These models typically produce feature maps at the following downsampling scales
+----------------------------------+------------------+
| cs3sedarknet_xdw | ✅ |
+----------------------------------+------------------+
| cspdarknet53 | ✅ |
+----------------------------------+------------------+
| cspresnet50 | ✅ |
+----------------------------------+------------------+
| cspresnet50d | ✅ |
Expand All @@ -107,6 +109,14 @@ These models typically produce feature maps at the following downsampling scales
+----------------------------------+------------------+
| cspresnext50 | ✅ |
+----------------------------------+------------------+
| darknet17 | ✅ |
+----------------------------------+------------------+
| darknet21 | ✅ |
+----------------------------------+------------------+
| darknet53 | ✅ |
+----------------------------------+------------------+
| darknetaa53 | ✅ |
+----------------------------------+------------------+
| densenet121 | |
+----------------------------------+------------------+
| densenet161 | |
Expand Down Expand Up @@ -189,14 +199,6 @@ These models typically produce feature maps at the following downsampling scales
+----------------------------------+------------------+
| eca_vovnet39b | |
+----------------------------------+------------------+
| ecaresnet101d | ✅ |
+----------------------------------+------------------+
| ecaresnet101d_pruned | ✅ |
+----------------------------------+------------------+
| ecaresnet200d | ✅ |
+----------------------------------+------------------+
| ecaresnet269d | ✅ |
+----------------------------------+------------------+
| ecaresnet26t | ✅ |
+----------------------------------+------------------+
| ecaresnet50d | ✅ |
Expand All @@ -205,6 +207,14 @@ These models typically produce feature maps at the following downsampling scales
+----------------------------------+------------------+
| ecaresnet50t | ✅ |
+----------------------------------+------------------+
| ecaresnet101d | ✅ |
+----------------------------------+------------------+
| ecaresnet101d_pruned | ✅ |
+----------------------------------+------------------+
| ecaresnet200d | ✅ |
+----------------------------------+------------------+
| ecaresnet269d | ✅ |
+----------------------------------+------------------+
| ecaresnetlight | ✅ |
+----------------------------------+------------------+
| ecaresnext26t_32x4d | ✅ |
Expand All @@ -213,10 +223,10 @@ These models typically produce feature maps at the following downsampling scales
+----------------------------------+------------------+
| efficientnet_b0 | ✅ |
+----------------------------------+------------------+
| efficientnet_b0_g16_evos | ✅ |
+----------------------------------+------------------+
| efficientnet_b0_g8_gn | ✅ |
+----------------------------------+------------------+
| efficientnet_b0_g16_evos | ✅ |
+----------------------------------+------------------+
| efficientnet_b0_gn | ✅ |
+----------------------------------+------------------+
| efficientnet_b1 | ✅ |
Expand Down Expand Up @@ -333,12 +343,12 @@ These models typically produce feature maps at the following downsampling scales
+----------------------------------+------------------+
| ghostnet_130 | |
+----------------------------------+------------------+
| ghostnetv2_050 | |
+----------------------------------+------------------+
| ghostnetv2_100 | |
+----------------------------------+------------------+
| ghostnetv2_130 | |
+----------------------------------+------------------+
| ghostnetv2_160 | |
+----------------------------------+------------------+
| halo2botnet50ts_256 | ✅ |
+----------------------------------+------------------+
| halonet26t | ✅ |
Expand Down Expand Up @@ -711,14 +721,14 @@ These models typically produce feature maps at the following downsampling scales
+----------------------------------+------------------+
| regnety_160 | ✅ |
+----------------------------------+------------------+
| regnety_1280 | ✅ |
+----------------------------------+------------------+
| regnety_2560 | ✅ |
+----------------------------------+------------------+
| regnety_320 | ✅ |
+----------------------------------+------------------+
| regnety_640 | ✅ |
+----------------------------------+------------------+
| regnety_1280 | ✅ |
+----------------------------------+------------------+
| regnety_2560 | ✅ |
+----------------------------------+------------------+
| regnetz_005 | ✅ |
+----------------------------------+------------------+
| regnetz_040 | ✅ |
Expand All @@ -733,12 +743,12 @@ These models typically produce feature maps at the following downsampling scales
+----------------------------------+------------------+
| regnetz_c16_evos | ✅ |
+----------------------------------+------------------+
| regnetz_d32 | ✅ |
+----------------------------------+------------------+
| regnetz_d8 | ✅ |
+----------------------------------+------------------+
| regnetz_d8_evos | ✅ |
+----------------------------------+------------------+
| regnetz_d32 | ✅ |
+----------------------------------+------------------+
| regnetz_e8 | ✅ |
+----------------------------------+------------------+
| repghostnet_050 | |
Expand Down Expand Up @@ -837,12 +847,12 @@ These models typically produce feature maps at the following downsampling scales
+----------------------------------+------------------+
| resnet50 | ✅ |
+----------------------------------+------------------+
| resnet50_gn | ✅ |
+----------------------------------+------------------+
| resnet50_clip | ✅ |
+----------------------------------+------------------+
| resnet50_clip_gap | ✅ |
+----------------------------------+------------------+
| resnet50_gn | ✅ |
+----------------------------------+------------------+
| resnet50_mlp | ✅ |
+----------------------------------+------------------+
| resnet50c | ✅ |
Expand Down Expand Up @@ -1001,6 +1011,8 @@ These models typically produce feature maps at the following downsampling scales
+----------------------------------+------------------+
| sebotnet33ts_256 | ✅ |
+----------------------------------+------------------+
| sedarknet21 | ✅ |
+----------------------------------+------------------+
| sehalonet33ts | ✅ |
+----------------------------------+------------------+
| selecsls42 | |
Expand Down Expand Up @@ -1045,14 +1057,6 @@ These models typically produce feature maps at the following downsampling scales
+----------------------------------+------------------+
| seresnetaa50d | ✅ |
+----------------------------------+------------------+
| seresnext101_32x4d | ✅ |
+----------------------------------+------------------+
| seresnext101_32x8d | ✅ |
+----------------------------------+------------------+
| seresnext101_64x4d | ✅ |
+----------------------------------+------------------+
| seresnext101d_32x8d | ✅ |
+----------------------------------+------------------+
| seresnext26d_32x4d | ✅ |
+----------------------------------+------------------+
| seresnext26t_32x4d | ✅ |
Expand All @@ -1061,6 +1065,14 @@ These models typically produce feature maps at the following downsampling scales
+----------------------------------+------------------+
| seresnext50_32x4d | ✅ |
+----------------------------------+------------------+
| seresnext101_32x4d | ✅ |
+----------------------------------+------------------+
| seresnext101_32x8d | ✅ |
+----------------------------------+------------------+
| seresnext101_64x4d | ✅ |
+----------------------------------+------------------+
| seresnext101d_32x8d | ✅ |
+----------------------------------+------------------+
| seresnextaa101d_32x8d | ✅ |
+----------------------------------+------------------+
| seresnextaa201d_32x8d | ✅ |
Expand Down Expand Up @@ -1163,6 +1175,22 @@ These models typically produce feature maps at the following downsampling scales
+----------------------------------+------------------+
| tinynet_e | ✅ |
+----------------------------------+------------------+
| vgg11 | |
+----------------------------------+------------------+
| vgg11_bn | |
+----------------------------------+------------------+
| vgg13 | |
+----------------------------------+------------------+
| vgg13_bn | |
+----------------------------------+------------------+
| vgg16 | |
+----------------------------------+------------------+
| vgg16_bn | |
+----------------------------------+------------------+
| vgg19 | |
+----------------------------------+------------------+
| vgg19_bn | |
+----------------------------------+------------------+
| vovnet39a | |
+----------------------------------+------------------+
| vovnet57a | |
Expand Down Expand Up @@ -1440,6 +1468,18 @@ Transformer-style models (e.g., Swin Transformer, ConvNeXt) typically produce fe
+------------------------------------+------------------+
| mvitv2_tiny | |
+------------------------------------+------------------+
| nest_base | |
+------------------------------------+------------------+
| nest_base_jx | |
+------------------------------------+------------------+
| nest_small | |
+------------------------------------+------------------+
| nest_small_jx | |
+------------------------------------+------------------+
| nest_tiny | |
+------------------------------------+------------------+
| nest_tiny_jx | |
+------------------------------------+------------------+
| nextvit_base | |
+------------------------------------+------------------+
| nextvit_large | |
Expand Down
Loading