8000 Change encoders · zyuh/segmentation_models.pytorch@0146fb1 · GitHub
[go: up one dir, main page]
8000 Skip to content

Commit 0146fb1

Browse files
qubvelqubvel
committed
Change encoders
1 parent 436c493 commit 0146fb1

File tree

1 file changed

+100
-33
lines changed

1 file changed

+100
-33
lines changed

‎README.md

Lines changed: 100 additions & 33 deletions
Original file line numberDiff line numberDiff line change
@@ -81,7 +81,7 @@ Congratulations! You are done! Now you can train your model with your favorite f
8181
#### Encoders <a name="encoders"></a>
8282

8383
<details>
84-
<summary>Table with ALL avaliable encoders (click to expand)</summary>
84+
<summary>ResNet</summary>
8585

8686
|Encoder |Weights |Params, M |
8787
|--------------------------------|:------------------------------:|:------------------------------:|
@@ -90,18 +90,42 @@ Congratulations! You are done! Now you can train your model with your favorite f
9090
|resnet50 |imagenet / ssl / swsl |23M |
9191
|resnet101 |imagenet |42M |
9292
|resnet152 |imagenet |58M |
93+
94+
</details>
95+
96+
<details>
97+
<summary>ResNeXt</summary>
98+
99+
|Encoder |Weights |Params, M |
100+
|--------------------------------|:------------------------------:|:------------------------------:|
93101
|resnext50_32x4d |imagenet / ssl / swsl |22M |
94102
|resnext101_32x4d |ssl / swsl |42M |
95103
|resnext101_32x8d |imagenet / instagram / ssl / swsl|86M |
96104
|resnext101_32x16d |instagram / ssl / swsl |191M |
97105
|resnext101_32x32d |instagram |466M |
98106
|resnext101_32x48d |instagram |826M |
107+
108+
</details>
109+
110+
<details>
111+
<summary>DPN</summary>
112+
113+
|Encoder |Weights |Params, M |
114+
|--------------------------------|:------------------------------:|:------------------------------:|
99115
|dpn68 |imagenet |11M |
100116
|dpn68b |imagenet+5k |11M |
101117
|dpn92 |imagenet+5k |34M |
102118
|dpn98 |imagenet |58M |
103119
|dpn107 |imagenet+5k |84M |
104120
|dpn131 |imagenet |76M |
121+
122+
</details>
123+
124+
<details>
125+
<summary>VGG</summary>
126+
127+
|Encoder |Weights |Params, M |
128+
|--------------------------------|:------------------------------:|:------------------------------:|
105129
|vgg11 |imagenet |9M |
106130
|vgg11_bn |imagenet |9M |
107131
|vgg13 |imagenet |9M |
@@ -110,18 +134,51 @@ Congratulations! You are done! Now you can train your model with your favorite f
110134
|vgg16_bn |imagenet |14M |
111135
|vgg19 |imagenet |20M |
112136
|vgg19_bn |imagenet |20M |
137+
138+
</details>
139+
140+
<details>
141+
<summary>SE-Net</summary>
142+
143+
|Encoder |Weights |Params, M |
144+
|--------------------------------|:------------------------------:|:------------------------------:|
113145
|senet154 |imagenet |113M |
114146
|se_resnet50 |imagenet |26M |
115147
|se_resnet101 |imagenet |47M |
116148
|se_resnet152 |imagenet |64M |
117149
|se_resnext50_32x4d |imagenet |25M |
118150
|se_resnext101_32x4d |imagenet |46M |
151+
152+
</details>
153+
154+
<details>
155+
<summary>DenseNet</summary>
156+
157+
|Encoder |Weights |Params, M |
158+
|--------------------------------|:------------------------------:|:------------------------------:|
119159
|densenet121 |imagenet |6M |
120160
|densenet169 |imagenet |12M |
121161
|densenet201 |imagenet |18M |
122162
|densenet161 |imagenet |26M |
163+
164+
</details>
165+
166+
<details>
167+
<summary>Inception</summary>
168+
169+
|Encoder |Weights |Params, M |
170+
|--------------------------------|:------------------------------:|:------------------------------:|
123171
|inceptionresnetv2 |imagenet / imagenet+background |54M |
124172
|inceptionv4 |imagenet / imagenet+background |41M |
173+
|xception |imagenet |22M |
174+
175+
</details>
176+
177+
<details>
178+
<summary>EfficientNet</summary>
179+
180+
|Encoder |Weights |Params, M |
181+
|--------------------------------|:------------------------------:|:------------------------------:|
125182
|efficientnet-b0 |imagenet |4M |
126183
|efficientnet-b1 |imagenet |6M |
127184
|efficientnet-b2 |imagenet |7M |
@@ -130,8 +187,6 @@ Congratulations! You are done! Now you can train your model with your favorite f
130187
|efficientnet-b5 |imagenet |28M |
131188
|efficientnet-b6 |imagenet |40M |
132189
|efficientnet-b7 |imagenet |63M |
133-
|mobilenet_v2 |imagenet |2M |
134-
|xception |imagenet |22M |
135190
|timm-efficientnet-b0 |imagenet / advprop / noisy-student|4M |
136191
|timm-efficientnet-b1 |imagenet / advprop / noisy-student|6M |
137192
|timm-efficientnet-b2 |imagenet / advprop / noisy-student|7M |
@@ -142,6 +197,23 @@ Congratulations! You are done! Now you can train your model with your favorite f
142197
|timm-efficientnet-b7 |imagenet / advprop / noisy-student|63M |
143198
|timm-efficientnet-b8 |imagenet / advprop |84M |
144199
|timm-efficientnet-l2 |noisy-student |474M |
200+
201+
</details>
202+
203+
<details>
204+
<summary>MobileNet</summary>
205+
206+
|Encoder |Weights |Params, M |
207+
|--------------------------------|:------------------------------:|:------------------------------:|
208+
|mobilenet_v2 |imagenet |2M |
209+
210+
</details>
211+
212+
<details>
213+
<summary>ResNeSt</summary>
214+
215+
|Encoder |Weights |Params, M |
216+
|--------------------------------|:------------------------------:|:------------------------------:|
145217
|timm-resnest14d |imagenet |8M |
146218
|timm-resnest26d |imagenet |15M |
147219
|timm-resnest50d |imagenet |25M |
@@ -150,13 +222,29 @@ Congratulations! You are done! Now you can train your model with your favorite f
150222
|timm-resnest269e |imagenet |108M |
151223
|timm-resnest50d_4s2x40d |imagenet |28M |
152224
|timm-resnest50d_1s4x24d |imagenet |23M |
225+
226+
</details>
227+
228+
<details>
229+
<summary>Res2Ne(X)t</summary>
230+
231+
|Encoder |Weights |Params, M |
232+
|--------------------------------|:------------------------------:|:------------------------------:|
153233
|timm-res2net50_26w_4s |imagenet |23M |
154234
|timm-res2net101_26w_4s |imagenet |43M |
155235
|timm-res2net50_26w_6s |imagenet |35M |
156236
|timm-res2net50_26w_8s |imagenet |46M |
157237
|timm-res2net50_48w_2s |imagenet |23M |
158238
|timm-res2net50_14w_8s |imagenet |23M |
159239
|timm-res2next50 |imagenet |22M |
240+
241+
</details>
242+
243+
<details>
244+
<summary>RegNet(x/y)</summary>
245+
246+
|Encoder |Weights |Params, M |
247+
|--------------------------------|:------------------------------:|:------------------------------:|
160248
|timm-regnetx_002 |imagenet |2M |
161249
|timm-regnetx_004 |imagenet |4M |
162250
|timm-regnetx_006 |imagenet |5M |
@@ -181,42 +269,21 @@ Congratulations! You are done! Now you can train your model with your favorite f
181269
|timm-regnety_120 |imagenet |49M |
182270
|timm-regnety_160 |imagenet |80M |
183271
|timm-regnety_320 |imagenet |141M |
184-
|timm-skresnet18 |imagenet |11M |
185-
|timm-skresnet34 |imagenet |21M |
186-
|timm-skresnext50_32x4d |imagenet |25M |
187-
188-
\* `ssl`, `wsl` - semi-supervised and weakly-supervised learning on ImageNet ([repo](https://github.com/facebookresearch/semi-supervised-ImageNet1K-models)).
189272

190273
</details>
191274

192-
Just commonly used encoders
275+
<details>
276+
<summary>SK-ResNe(X)t</summary>
193277

194278
|Encoder |Weights |Params, M |
195279
|--------------------------------|:------------------------------:|:------------------------------:|
196-
|resnet18 |imagenet / ssl / swsl |11M |
197-
|resnet34 |imagenet |21M |
198-
|resnet50 |imagenet / ssl / swsl |23M |
199-
|resnet101 |imagenet |42M |
200-
|resnext50_32x4d |imagenet / ssl / swsl |22M |
201-
|resnext101_32x4d |ssl / swsl |42M |
202-
|resnext101_32x8d |imagenet / instagram / ssl / swsl|86M |
203-
|senet154 |imagenet |113M |
204-
|se_resnext50_32x4d |imagenet |25M |
205-
|se_resnext101_32x4d |imagenet |46M |
206-
|densenet121 |imagenet |6M |
207-
|densenet169 |imagenet |12M |
208-
|densenet201 |imagenet |18M |
209-
|inceptionresnetv2 |imagenet / imagenet+background |54M |
210-
|inceptionv4 |imagenet / imagenet+background |41M |
211-
|mobilenet_v2 |imagenet |2M |
212-
|timm-efficientnet-b0 |imagenet / advprop / noisy-student|4M |
213-
|timm-efficientnet-b1 |imagenet / advprop / noisy-student|6M |
214-
|timm-efficientnet-b2 |imagenet / advprop / noisy-student|7M |
215-
|timm-efficientnet-b3 |imagenet / advprop / noisy-student|10M |
216-
|timm-efficientnet-b4 |imagenet / advprop / noisy-student|17M |
217-
|timm-efficientnet-b5 |imagenet / advprop / noisy-student|28M |
218-
|timm-efficientnet-b6 |imagenet / advprop / noisy-student|40M |
219-
|timm-efficientnet-b7 |imagenet / advprop / noisy-student|63M |
280+
|timm-skresnet18 |imagenet |11M |
281+
|timm-skresnet34 |imagenet |21M |
282+
|timm-skresnext50_32x4d |imagenet |25M |
283+
284+
</details>
285+
<br>
286+
\* `ssl`, `wsl` - semi-supervised and weakly-supervised learning on ImageNet ([repo](https://github.com/facebookresearch/semi-supervised-ImageNet1K-models)).
220287

221288

222289
### 🔁 Models API <a name="api"></a>

0 commit comments

Comments
 (0)
0