ScutterKey

PVA+RFCN

2

I read your paper and merge your base PVANet with RFCN. Have you tried using PVANet with RFCN? I found it hard to train and the accuracy(ignore_label=0) gets improved slowly....

mapping_names_resnets can not find

1

examples/contrib/quantization_aware_training/infer.py do not find mapping_names_resnets

reference code for the dataset generation method

5

Hi, does the reference code of dataset generation method release?

How to build the ios anakin from source ?

4

Thanks, I can only find the android build tutorial in your doc.

fix ios armv7 build issue https://github.com/PaddlePaddle/Anakin/issues/465

为什么CPU的速度能那么快？

32

谢谢。

Can the algorithm prune the mobilenet v1 or v2

1

Thanks!

I found your cfg(deeplab_resnet_v1_101_cityscapes_segmentation_dcn.yaml ) only use one gpu 0?

utils/image.py -> get_image ``` im, im_scale = resize(im, target_size, max_size, stride=config.network.IMAGE_STRIDE) im_tensor = transform(im, config.network.PIXEL_MEANS) processed_ims.append(im_tensor) im_info = [im_tensor.shape[2], im_tensor.shape[3], im_scale] # im_info is scaled and then padded ``` fpn/core/tester.py...