[FasterRCNN] update docs

examples/FasterRCNN/NOTES.md

@@ -21,10 +21,14 @@ This is a minimal implementation that simply contains these files:
 
 <p align="center"> <img src="https://user-images.githubusercontent.com/1381301/31527740-2f1b38ce-af84-11e7-8de1-628e90089826.png"> </p>
 
-3. Inference is not quite fast, because either you disable convolution autotune and end up with
+3. We use ROIAlign, and because of (3), `tf.image.crop_and_resize` is NOT ROIAlign.
+
+4. Inference is not quite fast, because either you disable convolution autotune and end up with
 	 a slow convolution algorithm, or you spend more time on autotune.
 	 This is a general problem of TensorFlow when running against variable-sized input.
 
-4. In Faster-RCNN, BatchNorm statistics are not supposed to be updated during fine-tuning.
+5. We only support single image per GPU for now.
+
+6. Because of (4), BatchNorm statistics are not supposed to be updated during fine-tuning.
 	 This specific kind of BatchNorm will need [my kernel](https://github.com/tensorflow/tensorflow/pull/12580)
 	 which is included since TF 1.4. If using an earlier version of TF, it will be either slow or wrong.

examples/FasterRCNN/README.md

 # Faster-RCNN on COCO
-This example aims to provide a minimal (<1000 lines) multi-GPU implementation of ResNet50-Faster-RCNN on COCO.
+This example aims to provide a minimal (1.2k lines) multi-GPU implementation of ResNet50-Faster-RCNN on COCO.
 
 ## Dependencies
-+ TensorFlow > 1.4.0 (use tf-nightly-gpu for now)
++ TensorFlow >= 1.4.0
 + Install [pycocotools](https://github.com/pdollar/coco/tree/master/PythonAPI/pycocotools), OpenCV.
 + Pre-trained [ResNet50 model](https://goo.gl/6XjK9V) from tensorpack model zoo.
 + COCO data. It assumes the following directory structure: