Jay Mahadeokar

You could use [pynetbuilder](https://github.com/jay-mahadeokar/pynetbuilder) for generating train_val.prototxt. Example: [imagenet training](https://github.com/jay-mahadeokar/pynetbuilder/tree/master/models/imagenet)

@KaimingHe do you think a link to another caffe re-implementation could be added in your Readme? This also provides a tool to generate other resnets in python.

It was renamed, please check this file for working version with imagenet: https://github.com/jay-mahadeokar/pynetbuilder/blob/master/app/imagenet/build_resnet.py

@weiliu89 thanks! Wow, did you finish training SSD on ResNet101 provided [here](https://github.com/KaimingHe/deep-residual-networks)? Can you please share what mAP you got on voc2007? Best I could get with ResNet50 is 70.4,...

@weiliu89 thanks! Your correct, the mAP improvements is marginal. Only advantage I can see is resnet_50 is much faster in terms of flops vs vggnet refer this [table](https://github.com/jay-mahadeokar/pynetbuilder/tree/master/models/voc2007_ssd#comparing-vgg-and-resnet-50-ssd-based-detection-networks), though I...

I believe so, I haven't benchmarked runtime on CPU though. FYI, I am using this [code](https://github.com/jay-mahadeokar/pynetbuilder/blob/master/netbuilder/tools/complexity.py#L31) to compute the flops.

@caocuong0306, I have open-sourced a tool to generate residual networks and object detection caffe networks here: https://github.com/jay-mahadeokar/pynetbuilder. Example: [resnet training on imagenet](https://github.com/jay-mahadeokar/pynetbuilder/tree/master/models/imagenet) , [object detection using resnet + ssd](https://github.com/jay-mahadeokar/pynetbuilder/tree/master/models/voc2007_ssd) Could...

@kaishijeng I believe the complexity of squeezenet in terms of flops is ~800 Million (though not sure, need to run it through complexity module) and the corresponding top1 accuracy on...

Please refer this [table](https://github.com/jay-mahadeokar/pynetbuilder/tree/master/models/voc2007_ssd#comparing-vgg-and-resnet-50-ssd-based-detection-networks) for ssd+vgg16 and ssd+resnet50. I have also shared the caffemodel.. This [table](https://github.com/jay-mahadeokar/pynetbuilder/tree/master/models/imagenet#basic-residual-network-results) also compares resnet 50 and resnet_50_1by2. Though I havent yet added model files object...

I haven't done thorough benchmarking on cpu, since I only tested validation set on gpu machines. but I guess that should be true! I will run it on CPU and...