Abhishek Agrawal

@lucasb-eyer So what do you think, in what cases I should not initialize the head with zeros? Can you think of a problem statement where I should not initialize with...

@ersheng-ai How do I run validation on coco ?

@WongKinYiu It seems.. The concept of splitting the input into two parts is exactly same as group convolution ? Please explain If I am wrong.

@hitbuyi I have successfully export DCNv2 to ONNX and TensorRT. The trick is - use mmcv's ModulatedDeformConv2D instead of this repo. Check this for more details - https://medium.com/@abhiagwl4262/yolact-pyotorch-to-onnx-tensorrt-9e6a37f6b974

@ArtyZe Same observation. Came up with any idea ?

I am still getting the following error /mnt/abhi/tensorflow/tensorflow/contrib/avro/BUILD:13:1: no such package '@avro_archive//': Prefix avro-cpp-1.8.0 was given, but not found in the archive and referenced by '//tensorflow/contrib/avro:example_h'

when I try to add them up in a single line of list comprehension, I get an error stating the sizes for gradients are different. This is somewhat expected since...

@mrharicot I changed the following functions. def gradient_x(self, img): paddings = tf.constant([[0,0],[0,0],[1,0],[0,0]]) img = tf.pad(img, paddings,'CONSTANT') gx = img[:,:,:-1,:] - img[:,:,1:,:] return gx def gradient_y(self, img): paddings = tf.constant([[0,0],[1,0],[0,0],[0,0]]) img...

@mrharicot Will there be any issue if I use "tf.image.ssim" instead of self.SSIM ???

Please follow this article for OnnxRuntime and TensorRT deployment. https://medium.com/@abhiagwl4262/yolact-pyotorch-to-onnx-tensorrt-9e6a37f6b974