Upd info about several models (ResNeXt-101 (32x4d), DPN-98, RegNetX-1…

…6GF, ResNeSt(A)-50, SE-ResNet-152)
osmr · Mar 13, 2021 · 537fd0b · 537fd0b
1 parent 076bbce
commit 537fd0b
Show file tree

Hide file tree

Showing 8 changed files with 40 additions and 40 deletions.
diff --git a/chainer_/README.md b/chainer_/README.md
@@ -208,7 +208,7 @@ Some remarks:
 | ResNeXt-26 (32x2d) | 26.32 | 8.49 | 9,924,136 | 1,461.06M | Converted from GL model ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.373/resnext26_32x2d-0849-58d86996.npz.log)) |
 | ResNeXt-26 (32x4d) | 23.91 | 7.19 | 15,389,480 | 2,488.07M | Converted from GL model ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.332/resnext26_32x4d-0719-62ca5090.npz.log)) |
 | ResNeXt-50 (32x4d) | 20.84 | 5.46 | 25,028,904 | 4,255.86M | Converted from GL model ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.498/resnext50_32x4d-0546-d1e20cc5.npz.log)) |
-| ResNeXt-101 (32x4d) | 18.54 | 4.20 | 44,177,704 | 8,003.45M | Converted from GL model ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.510/resnext101_32x4d-0420-a8e1396e.npz.log)) |
+| ResNeXt-101 (32x4d) | 18.51 | 4.16 | 44,177,704 | 8,003.45M | Converted from GL model ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.530/resnext101_32x4d-0416-ec66f27f.npz.log)) |
 | ResNeXt-101 (64x4d) | 19.31 | 4.85 | 83,455,272 | 15,500.27M | From [dmlc/gluon-cv] ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.417/resnext101_64x4d-0485-b3c1a220.npz.log)) |
 | SE-ResNet-10 | 31.41 | 11.70 | 5,463,332 | 894.27M | Converted from GL model ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.486/seresnet10-1170-2b3424cb.npz.log)) |
 | SE-ResNet-18 | 27.96 | 9.23 | 11,778,592 | 1,820.88M | Converted from GL model ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.355/seresnet18-0923-b0931abe.npz.log)) |
@@ -219,7 +219,7 @@ Some remarks:
 | SE-ResNet-50b | 20.56 | 5.30 | 28,088,024 | 4,115.78M | Converted from GL model ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.387/seresnet50b-0530-1ac3bf50.npz.log)) |
 | SE-ResNet-101 | 19.59 | 4.74 | 49,326,872 | 7,602.76M | Converted from GL model ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.508/seresnet101-0474-828aaed6.npz.log)) |
 | SE-ResNet-101b | 19.48 | 4.63 | 49,326,872 | 7,839.75M | Converted from GL model ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.460/seresnet101b-0463-97cc55c3.npz.log)) |
-| SE-ResNet-152 | 21.46 | 5.77 | 66,821,848 | 11,328.52M | From [Cadene/pretrained...pytorch] ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.24/seresnet152-0577-a089ba52.npz.log)) |
+| SE-ResNet-152 | 19.94 | 4.99 | 66,821,848 | 11,328.52M | Converted from GL model ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.530/seresnet152-0499-9be4deeb.npz.log)) |
 | SE-PreResNet-10 | 33.63 | 13.11 | 5,461,668 | 894.42M | Converted from GL model ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.377/sepreresnet10-1311-5e38607c.npz.log)) |
 | SE-PreResNet-18 | 27.72 | 9.39 | 11,776,928 | 1,821.03M | Converted from GL model ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.380/sepreresnet18-0939-a78ded77.npz.log)) |
 | SE-PreResNet-BC-26b | 22.93 | 6.38 | 17,388,424 | 2,357.07M | Converted from GL model ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.399/sepreresnetbc26b-0638-e8393574.npz.log)) |
@@ -234,7 +234,7 @@ Some remarks:
 | ResNeSt(A)-BC-14 | 22.28 | 6.33 | 10,611,688 | 2,767.37M | Converted from GL model ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.493/resnestabc14-0633-a76f3b83.npz.log)) |
 | ResNeSt(A)-18 | 23.43 | 6.94 | 12,763,784 | 2,587.50M | Converted from GL model ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.489/resnesta18-0694-4ecaf0b7.npz.log)) |
 | ResNeSt(A)-BC-26 | 19.59 | 4.69 | 17,069,448 | 3,646.57M | Converted from GL model ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.495/resnestabc26-0469-14cb6ca5.npz.log)) |
-| ResNeSt(A)-50 | 18.88 | 4.52 | 27,483,240 | 5,403.11M | From [dmlc/gluon-cv] ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.465/resnesta50-0452-8cfddfdb.npz.log)) |
+| ResNeSt(A)-50 | 18.98 | 4.40 | 27,483,240 | 5,403.11M | Converted from GL model ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.530/resnesta50-0440-f3a4a191.npz.log)) |
 | ResNeSt(A)-101 | 17.73 | 4.00 | 48,275,016 | 10,247.88M | From [dmlc/gluon-cv] ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.465/resnesta101-0400-bd2efb42.npz.log)) |
 | ResNeSt(A)-200 | 16.81 | 3.39 | 70,201,544 | 22,857.88M | From [dmlc/gluon-cv] ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.465/resnesta200-0339-b521427b.npz.log)) |
 | ResNeSt(A)-269 | 16.40 | 3.36 | 110,929,480 | 46,012.95M | From [dmlc/gluon-cv] ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.465/resnesta269-0336-933dbe64.npz.log)) |
@@ -256,7 +256,7 @@ Some remarks:
 | RegNetX-6.4GF | 19.64 | 4.81 | 26,209,256 | 6,491.01M | Converted from GL model ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.513/regnetx064-0481-06504640.npz.log)) |
 | RegNetX-8.0GF | 19.62 | 4.67 | 39,572,648 | 8,017.94M | Converted from GL model ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.515/regnetx080-0467-3618e40b.npz.log)) |
 | RegNetX-12GF | 20.29 | 5.24 | 46,106,056 | 12,124.22M | From [rwightman/pyt...models] ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.473/regnetx120-0524-50b59d58.npz.log)) |
-| RegNetX-16GF | 19.97 | 5.04 | 54,278,536 | 15,986.64M | From [rwightman/pyt...models] ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.473/regnetx160-0504-10fd49b1.npz.log)) |
+| RegNetX-16GF | 19.62 | 4.74 | 54,278,536 | 15,986.64M | Converted from GL model ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.530/regnetx160-0474-ca36a68b.npz.log)) |
 | RegNetX-32GF | 19.57 | 4.87 | 107,811,560 | 31,790.24M | From [rwightman/pyt...models] ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.473/regnetx320-0487-caa9632a.npz.log)) |
 | RegNetY-200MF | 28.50 | 9.55 | 3,162,996 | 203.99M | Converted from GL model ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.476/regnety002-0955-5ba3e62c.npz.log)) |
 | RegNetY-400MF | 24.84 | 7.52 | 4,344,144 | 410.35M | Converted from GL model ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.481/regnety004-0752-e30b7c27.npz.log)) |
@@ -289,7 +289,7 @@ Some remarks:
 | DRN-D-54 | 22.08 | 6.23 | 35,809,176 | 28,547.38M | From [fyu/drn] ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.116/drnd54-0623-31e8eeb8.npz.log)) |
 | DRN-D-105 | 21.32 | 5.84 | 54,801,304 | 43,442.43M | From [fyu/drn] ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.116/drnd105-0584-c0d7657b.npz.log)) |
 | DPN-68 | 22.93 | 6.56 | 12,611,602 | 2,351.84M | Converted from GL model ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.310/dpn68-0656-bf9b72e9.npz.log)) |
-| DPN-98 | 20.80 | 5.53 | 61,570,728 | 11,716.51M | From [Cadene/pretrained...pytorch] ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.34/dpn98-0553-9cd57335.npz.log)) |
+| DPN-98 | 19.70 | 5.15 | 61,570,728 | 11,716.51M | Converted from GL model ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.530/dpn98-0515-fbb00476.npz.log)) |
 | DPN-131 | 20.04 | 5.23 | 79,254,504 | 16,076.15M | From [Cadene/pretrained...pytorch] ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.34/dpn131-0523-e3721599.npz.log)) |
 | DarkNet Tiny | 40.33 | 17.46 | 1,042,104 | 500.85M | Converted from GL model ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.69/darknet_tiny-1746-b04fa463.npz.log)) |
 | DarkNet Ref | 38.09 | 16.71 | 7,319,416 | 367.59M | Converted from GL model ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.64/darknet_ref-1671-b2d5721f.npz.log)) |

diff --git a/chainer_/chainercv2/models/model_store.py b/chainer_/chainercv2/models/model_store.py
@@ -73,7 +73,7 @@
     ('resnext26_32x2d', '0849', '58d86996a5e83efe1f49513b7ef5e7803f4830b6', 'v0.0.373'),
     ('resnext26_32x4d', '0719', '62ca50907121ceee5aada95eda7398dca69928cb', 'v0.0.332'),
     ('resnext50_32x4d', '0546', 'd1e20cc5a1c88a131033e2b1c4f4e06d28084c73', 'v0.0.498'),
-    ('resnext101_32x4d', '0420', 'a8e1396e12e2c7a79d32ebfef7e255f989798a77', 'v0.0.510'),
+    ('resnext101_32x4d', '0416', 'ec66f27ffcc27337c5b7d74f68cd3b0b466fb22c', 'v0.0.530'),
     ('resnext101_64x4d', '0485', 'b3c1a22070960a72dc9e46588fc27b57b1ff98e0', 'v0.0.417'),
     ('seresnet10', '1170', '2b3424cb76b192830ac243df3f5d60f2a0167e9c', 'v0.0.486'),
     ('seresnet18', '0923', 'b0931abe71a56c02e9c1012d34b99e469602eacc', 'v0.0.355'),
@@ -84,7 +84,7 @@
     ('seresnet50b', '0530', '1ac3bf504713faa817e7cf1c5e8e37fb4368a883', 'v0.0.387'),
     ('seresnet101', '0474', '828aaed6faa46abadbe82f5d8386acf5ebdc9430', 'v0.0.508'),
     ('seresnet101b', '0463', '97cc55c354860719b9c206d6808e2ebd7a019bf9', 'v0.0.460'),
-    ('seresnet152', '0577', 'a089ba52930e9949313b9fba00a1b2e6e68f6ea4', 'v0.0.24'),
+    ('seresnet152', '0499', '9be4deeb22052eb14d8e456483e9af7d2d8284a6', 'v0.0.530'),
     ('sepreresnet10', '1311', '5e38607cfb4971dbd01d6b98058975bbeb6ff61a', 'v0.0.377'),
     ('sepreresnet18', '0939', 'a78ded77bd89ce527d606ebaffb75f0cd2258b2e', 'v0.0.380'),
     ('sepreresnetbc26b', '0638', 'e8393574e457106c4963f7979d91b67dc7554239', 'v0.0.399'),
@@ -99,7 +99,7 @@
     ('resnestabc14', '0633', 'a76f3b83d2ecab7ee278000f6d205ddae648e0bf', 'v0.0.493'),
     ('resnesta18', '0694', '4ecaf0b7ce07224b4eb1972562ec8aa39e481796', 'v0.0.489'),
     ('resnestabc26', '0469', '14cb6ca540ef6cb8feb35199b015e41d69fb5df6', 'v0.0.495'),
-    ('resnesta50', '0452', '8cfddfdbc6bee0d0d0c6770e05710d4866513371', 'v0.0.465'),
+    ('resnesta50', '0440', 'f3a4a1914ea38d23a50c99b4e2e89378ab6e73b5', 'v0.0.530'),
     ('resnesta101', '0400', 'bd2efb4282ecce70dfb7985a2fbda9a0ee11a955', 'v0.0.465'),
     ('resnesta200', '0339', 'b521427b8e168879455118c21ed4d99f2d37a682', 'v0.0.465'),
     ('resnesta269', '0336', '933dbe64fb27a174752b77980bad3c545582c5e8', 'v0.0.465'),
@@ -121,7 +121,7 @@
     ('regnetx064', '0481', '06504640aee55e94949bec65fe8e776199a28e06', 'v0.0.513'),
     ('regnetx080', '0467', '3618e40bed12247e0f27027f09adecceafc514eb', 'v0.0.515'),
     ('regnetx120', '0524', '50b59d580d73bc87c063f2fdd2e2d9a77db25971', 'v0.0.473'),
-    ('regnetx160', '0504', '10fd49b109ceacde22284ecff5eb9e682a5345ec', 'v0.0.473'),
+    ('regnetx160', '0474', 'ca36a68bce3526c6ed54e50d312af1e004d09aae', 'v0.0.530'),
     ('regnetx320', '0487', 'caa9632aecc91b1bcfb2562dc2f7106c493e7934', 'v0.0.473'),
     ('regnety002', '0955', '5ba3e62c853d6216aadd189c49cfa2dc0d9fd871', 'v0.0.476'),
     ('regnety004', '0752', 'e30b7c274003dad0f2fa5c8e16e2e34992acc3ea', 'v0.0.481'),
@@ -154,7 +154,7 @@
     ('drnd54', '0623', '31e8eeb88bdbb07d8613a16471c8c5bd67ae823a', 'v0.0.116'),
     ('drnd105', '0584', 'c0d7657b2d3c4cf7d97ff407cd50dda5d1bd1880', 'v0.0.116'),
     ('dpn68', '0656', 'bf9b72e9749da4c6ee5a544639f78ac7fa85f7ce', 'v0.0.310'),
-    ('dpn98', '0553', '9cd5733573f7a99062d16cd8850bb82d684704bb', 'v0.0.34'),
+    ('dpn98', '0515', 'fbb00476a16114928f3db0eaef63c947536efabd', 'v0.0.530'),
     ('dpn131', '0523', 'e37215991fa7e9f49245843d53de63ef1717f293', 'v0.0.34'),
     ('darknet_tiny', '1746', 'b04fa46318a78e977aa5a117786968d98d325871', 'v0.0.69'),
     ('darknet_ref', '1671', 'b2d5721f3a5f6f05cc785d57ff7a63fe82f6325e', 'v0.0.64'),

diff --git a/gluon/README.md b/gluon/README.md
@@ -222,7 +222,7 @@ bottleneck block. Respectively a network without b-suffix has the stride in the
 | ResNeXt-26 (32x2d) | 26.34 | 8.50 | 9,924,136 | 1,461.06M | Training ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.373/resnext26_32x2d-0850-a1fb4451.params.log)) |
 | ResNeXt-26 (32x4d) | 23.93 | 7.21 | 15,389,480 | 2,488.07M | Training ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.332/resnext26_32x4d-0721-5264d7ef.params.log)) |
 | ResNeXt-50 (32x4d) | 20.84 | 5.45 | 25,028,904 | 4,255.86M | Training ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.498/resnext50_32x4d-0545-748e53b3.params.log)) |
-| ResNeXt-101 (32x4d) | 18.53 | 4.19 | 44,177,704 | 8,003.45M | Training ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.510/resnext101_32x4d-0419-fca3fb03.params.log)) |
+| ResNeXt-101 (32x4d) | 18.46 | 4.18 | 44,177,704 | 8,003.45M | Training ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.530/resnext101_32x4d-0418-c07faa66.params.log)) |
 | ResNeXt-101 (64x4d) | 19.28 | 4.83 | 83,455,272 | 15,500.27M | From [dmlc/gluon-cv] ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.417/resnext101_64x4d-0483-44b79943.params.log)) |
 | SE-ResNet-10 | 31.36 | 11.69 | 5,463,332 | 894.08M | Training ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.486/seresnet10-1169-675d4b5b.params.log)) |
 | SE-ResNet-18 | 27.95 | 9.20 | 11,778,592 | 1,820.51M | Training ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.355/seresnet18-0920-85a6b1da.params.log)) |
@@ -233,7 +233,7 @@ bottleneck block. Respectively a network without b-suffix has the stride in the
 | SE-ResNet-50b | 20.58 | 5.33 | 28,088,024 | 4,113.02M | Training ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.387/seresnet50b-0533-0d8f0d23.params.log)) |
 | SE-ResNet-101 | 19.53 | 4.75 | 49,326,872 | 7,602.76M | Training ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.508/seresnet101-0475-540a33f4.params.log)) |
 | SE-ResNet-101b | 19.46 | 4.62 | 49,326,872 | 7,835.29M | Training ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.460/seresnet101b-0462-59fae71a.params.log)) |
-| SE-ResNet-152 | 21.48 | 5.77 | 66,821,848 | 11,328.52M | From [Cadene/pretrained...pytorch] ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.11/seresnet152-0577-de6f099d.params.log)) |
+| SE-ResNet-152 | 19.95 | 4.98 | 66,821,848 | 11,328.52M | Training ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.530/seresnet152-0498-299b377e.params.log)) |
 | SE-PreResNet-10 | 33.60 | 13.06 | 5,461,668 | 894.23M | Training ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.377/sepreresnet10-1306-cbdd1053.params.log)) |
 | SE-PreResNet-18 | 27.67 | 9.38 | 11,776,928 | 1,820.66M | Training ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.380/sepreresnet18-0938-f9645ed3.params.log)) |
 | SE-PreResNet-BC-26b | 22.95 | 6.36 | 17,388,424 | 2,355.57M | Training ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.399/sepreresnetbc26b-0636-33c94c9d.params.log)) |
@@ -248,7 +248,7 @@ bottleneck block. Respectively a network without b-suffix has the stride in the
 | ResNeSt(A)-BC-14 | 22.27 | 6.34 | 10,611,688 | 2,766.86M | Training ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.493/resnestabc14-0634-4b0cbe8c.params.log)) |
 | ResNeSt(A)-18 | 23.43 | 6.89 | 12,763,784 | 2,587.11M | Training ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.489/resnesta18-0689-8f37b692.params.log)) |
 | ResNeSt(A)-BC-26 | 19.57 | 4.70 | 17,069,448 | 3,645.87M | Training ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.495/resnestabc26-0470-f88d49d7.params.log)) |
-| ResNeSt(A)-50 | 18.89 | 4.51 | 27,483,240 | 5,402.09M | From [dmlc/gluon-cv] ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.465/resnesta50-0451-445a013a.params.log)) |
+| ResNeSt(A)-50 | 18.99 | 4.38 | 27,483,240 | 5,402.09M | Training ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.530/resnesta50-0438-90b0faaf.params.log)) |
 | ResNeSt(A)-101 | 17.74 | 3.99 | 48,275,016 | 10,246.42M | From [dmlc/gluon-cv] ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.465/resnesta101-0399-ab6c6f89.params.log)) |
 | ResNeSt(A)-200 | 16.78 | 3.40 | 70,201,544 | 22,854.22M | From [dmlc/gluon-cv] ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.465/resnesta200-0340-3bd1f0c8.params.log)) |
 | ResNeSt(A)-269 | 16.38 | 3.36 | 110,929,480 | 46,005.88M | From [dmlc/gluon-cv] ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.465/resnesta269-0336-8333862a.params.log)) |
@@ -276,7 +276,7 @@ bottleneck block. Respectively a network without b-suffix has the stride in the
 | RegNetX-6.4GF | 19.69 | 4.84 | 26,209,256 | 6,490.97M | Training ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.513/regnetx064-0484-94225eb7.params.log)) |
 | RegNetX-8.0GF | 19.62 | 4.66 | 39,572,648 | 8,017.90M | Training ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.515/regnetx080-0466-4086910a.params.log)) |
 | RegNetX-12GF | 20.34 | 5.22 | 46,106,056 | 12,124.16M | From [rwightman/pyt...models] ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.473/regnetx120-0522-22c6c138.params.log)) |
-| RegNetX-16GF | 20.01 | 5.05 | 54,278,536 | 15,986.59M | From [rwightman/pyt...models] ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.473/regnetx160-0505-f2dac945.params.log)) |
+| RegNetX-16GF | 19.60 | 4.74 | 54,278,536 | 15,986.59M | Training ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.530/regnetx160-0474-a1174530.params.log)) |
 | RegNetX-32GF | 19.57 | 4.89 | 107,811,560 | 31,790.18M | From [rwightman/pyt...models] ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.473/regnetx320-0489-80ef5db7.params.log)) |
 | RegNetY-200MF | 28.50 | 9.53 | 3,162,996 | 203.80M | Training ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.476/regnety002-0953-b37fcac0.params.log)) |
 | RegNetY-400MF | 24.85 | 7.47 | 4,344,144 | 409.95M | Training ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.481/regnety004-0747-5626bdf4.params.log)) |
@@ -310,7 +310,7 @@ bottleneck block. Respectively a network without b-suffix has the stride in the
 | DRN-D-54 | 22.05 | 6.27 | 35,809,176 | 28,547.38M | From [fyu/drn] ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.116/drnd54-0627-87d44c87.params.log)) |
 | DRN-D-105 | 21.31 | 5.81 | 54,801,304 | 43,442.43M | From [fyu/drn] ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.116/drnd105-0581-ab12d662.params.log)) |
 | DPN-68 | 22.87 | 6.58 | 12,611,602 | 2,351.84M | Training ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.310/dpn68-0658-07251919.params.log)) |
-| DPN-98 | 20.23 | 5.28 | 61,570,728 | 11,716.51M | From [Cadene/pretrained...pytorch] ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.17/dpn98-0528-fa5d6fca.params.log)) |
+| DPN-98 | 19.68 | 5.13 | 61,570,728 | 11,716.51M | Training ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.530/dpn98-0513-6ec5d288.params.log)) |
 | DPN-131 | 20.03 | 5.22 | 79,254,504 | 16,076.15M | From [Cadene/pretrained...pytorch] ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.17/dpn131-0522-35ac2f82.params.log)) |
 | DarkNet Tiny | 40.31 | 17.46 | 1,042,104 | 500.85M | Training ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.69/darknet_tiny-1746-16501793.params.log)) |
 | DarkNet Ref | 38.00 | 16.68 | 7,319,416 | 367.59M | Training ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.64/darknet_ref-1668-3011b4e1.params.log)) |