[MaskRCNN] add stride1x1, update table

examples/FasterRCNN/README.md

@@ -62,16 +62,17 @@ Evaluation or prediction will need the same config used during training.
 These models are trained with different configurations on trainval35k and evaluated on minival using mAP@IoU=0.50:0.95.
 MaskRCNN results contain both bbox and segm mAP.
 
-|Backbone|`FASTRCNN_BATCH`|resolution |schedule|mAP (bbox/segm)|Time          |
-|   -    |    -           |    -      |   -    |   -           |   -          |
-|R50-C4  |64              |(600, 1024)|280k    |33.1           |18h on 8 V100s|
-|R50-C4  |512             |(800, 1333)|360k    |36.6           |49h on 8 V100s|
-|R50-FPN |512             |(800, 1333)|360k    |37.5           |28h on 8 V100s|
-|R50-C4  |256             |(800, 1333)|280k    |36.8/32.1      |39h on 8 P100s|
-|R50-C4  |512             |(800, 1333)|360k    |37.8/33.1      |51h on 8 V100s|
-|R50-FPN |512             |(800, 1333)|360k    |38.1/34.9      |38h on 8 V100s|
-|R101-C4 |512             |(800, 1333)|280k    |40.1/34.4      |70h on 8 P100s|
-|R101-C4 |512             |(800, 1333)|360k    |40.8/35.1      |63h on 8 V100s|
+ | Backbone | mAP<br/>(box/mask) | Detectron mAP <br/> (box/mask) | Time           | Configurations <br/> (click to expand)                                                                                                                                              |
+ | -        | -                  | -                              | -              | -                                                                                                                                                                                   |
+ | R50-C4   | 33.1               |                                | 18h on 8 V100s | <details><summary>super quick</summary>`MODE_MASK=False FRCNN.BATCH_PER_IM=64 PREPROC.SHORT_EDGE_SIZE=600 PREPROC.MAX_SIZE=1024 TRAIN.LR_SCHEDULE=[150000,230000,280000]` </details> |
+ | R50-C4   | 36.6               | 36.5                           | 49h on 8 V100s | <details><summary>standard</summary>`MODE_MASK=False` </details>                                                                                                                    |
+ | R50-FPN  | 37.5               | 37.9<sup>[1](#ft1)</sup>       | 28h on 8 V100s | <details><summary>standard</summary>`MODE_MASK=False MODE_FPN=True` </details>                                                                                                      |
+ | R50-C4   | 36.8/32.1          |                                | 39h on 8 P100s | <details><summary>quick</summary>`MODE_MASK=True FRCNN.BATCH_PER_IM=256 TRAIN.LR_SCHEDULE=[150000,230000,280000]` </details>                                                         |
+ | R50-C4   | 37.8/33.1          | 37.8/32.8                      | 51h on 8 V100s | <details><summary>standard</summary>`MODE_MASK=True` </details>                                                                                                                     |
+ | R50-FPN  | 38.1/34.9          | 38.6/34.5<sup>[1](#ft1)</sup>  | 38h on 8 V100s | <details><summary>standard</summary>`MODE_MASK=True MODE_FPN=True` </details>                                                                                                       |
+ | R101-C4  | 40.8/35.1          |                                | 63h on 8 V100s | <details><summary>standard</summary>`MODE_MASK=True BACKBONE.RESNET_NUM_BLOCK=[3,4,23,3]` </details>                                                                                |
+ 
+ <a id="ft1">1</a>: Slightly different configurations.
 
 The two R50-C4 360k models have the same configuration __and mAP__
 as the `R50-C4-2x` entries in

examples/FasterRCNN/basemodel.py

@@ -85,8 +85,14 @@ def resnet_shortcut(l, n_out, stride, activation=tf.identity):
 
 
 def resnet_bottleneck(l, ch_out, stride):
-    l, shortcut = l, l
-    l = Conv2D('conv1', l, ch_out, 1, activation=BNReLU)
+    shortcut = l
+    if cfg.BACKBONE.STRIDE_1X1:
+        if stride == 2:
+            l = l[:, :, :-1, :-1]
+        l = Conv2D('conv1', l, ch_out, 1, strides=stride, activation=BNReLU)
+        l = Conv2D('conv2', l, ch_out, 3, strides=1, activation=BNReLU)
+    else:
+        l = Conv2D('conv1', l, ch_out, 1, strides=1, activation=BNReLU)
         if stride == 2:
             l = tf.pad(l, [[0, 0], [0, 0], maybe_reverse_pad(0, 1), maybe_reverse_pad(0, 1)])
             l = Conv2D('conv2', l, ch_out, 3, strides=2, activation=BNReLU, padding='VALID')

examples/FasterRCNN/config.py

@@ -72,6 +72,7 @@ _C.BACKBONE.NORM = 'FreezeBN'  # options: FreezeBN, SyncBN
 # We will eventually switch to TF_PAD_MODE=False.
 # See https://github.com/tensorflow/tensorflow/issues/18213
 _C.BACKBONE.TF_PAD_MODE = True
+_C.BACKBONE.STRIDE_1X1 = False  # True for MSRA models
 
 # schedule -----------------------
 # The schedule and learning rate here is defined for a total batch size of 8.
@@ -82,7 +83,6 @@ _C.TRAIN.BASE_LR = 1e-2
 _C.TRAIN.WARMUP = 1000    # in steps
 _C.TRAIN.STEPS_PER_EPOCH = 500
 # LR_SCHEDULE = [120000, 160000, 180000]  # "1x" schedule in detectron
-# LR_SCHEDULE = [150000, 230000, 280000]  # roughly a "1.5x" schedule
 _C.TRAIN.LR_SCHEDULE = [240000, 320000, 360000]    # "2x" schedule in detectron
 
 # preprocessing --------------------