habana_frameworks.mediapipe.fn.Mult

Class:

• habana_frameworks.mediapipe.fn.Mult(**kwargs)

Define graph call:

• __call__(input1, input2)

Parameter:

• input1 - First input tensor to operator. Supported dimensions: minimum = 1, maximum = 5. Supported data types: INT16, INT32, FLOAT16, BFLOAT16, FLOAT32.

• input2 - Second input tensor to operator. Supported dimensions: minimum = 1, maximum = 5. Supported data types: INT16, INT32, FLOAT16, BFLOAT16, FLOAT32.

Description:

Resultant tensor formed from the multiplication of the two operands element-wise. This operator performs element-wise multiplication and supports Broadcasting. Computes output as: output = (input1 * input2), element-wise.

Supported backend:

• HPU

Keyword Arguments

kwargs	Description
dtype	Output data type. Type: habana_frameworks.mediapipe.media_types.dtype Default: FLOAT32 Optional: yes Supported data types: INT32 BFLOAT16 FLOAT32

Note

1. All input/output tensors must be of the same data type and must have the same dimensionality except in broadcast support where dimensionality can be different.

2. This operator is agnostic to the data layout.

Example: Mult Operator

The following code snippet shows usage of Mult operator:

from habana_frameworks.mediapipe import fn
from habana_frameworks.mediapipe.mediapipe import MediaPipe
from habana_frameworks.mediapipe.media_types import imgtype as it
from habana_frameworks.mediapipe.media_types import dtype as dt
import matplotlib.pyplot as plt
from numpy.random import RandomState
import numpy as np

# Create media pipeline derived class
class myMediaPipe(MediaPipe):
    def __init__(self, device, dir, queue_depth, batch_size, img_h, img_w, img_c):
        super(
            myMediaPipe,
            self).__init__(
            device,
            queue_depth,
            batch_size,
            self.__class__.__name__)

        self.input = fn.ReadImageDatasetFromDir(shuffle=False,
                                                dir=dir,
                                                format="jpg")

        rng = RandomState(seed=100)

        self.input_data = rng.rand(img_c,
                                   img_w,
                                   img_h,
                                   batch_size)

        self.input_data = self.input_data.astype(np.float32)

        self.input_node = fn.MediaConst(data=self.input_data,
                                        shape=[img_w, img_h, img_c, batch_size],
                                        dtype=dt.FLOAT32)

        self.cast = fn.Cast(dtype=dt.FLOAT32, round_mode=0)

        # WHCN
        self.decode = fn.ImageDecoder(device="hpu",
                                      output_format=it.RGB_P,
                                      resize=[img_w, img_h])

        self.mult = fn.Mult(dtype=dt.FLOAT32)

        # WHCN -> CWHN
        self.transpose = fn.Transpose(permutation=[2, 0, 1, 3],
                                      tensorDim=4,
                                      dtype=dt.FLOAT32)

    def definegraph(self):
        images, labels = self.input()
        images = self.decode(images)
        images1 = self.cast(images)
        images2 = self.input_node()
        images = self.mult(images1, images2)
        images = self.transpose(images)
        return images, labels

def main():
    batch_size = 6
    img_width = 200
    img_height = 200
    img_channel = 3
    img_dir = "/path/to/images"
    queue_depth = 2
    columns = 3

    # Create media pipeline object
    pipe = myMediaPipe('hpu', img_dir, queue_depth, batch_size,
                        img_height, img_width, img_channel)

    # Build media pipeline
    pipe.build()

    # Initialize media pipeline iterator
    pipe.iter_init()

    # Run media pipeline
    images, labels = pipe.run()

    # Copy data to host from device as numpy array
    images = images.as_cpu().as_nparray()
    labels = labels.as_cpu().as_nparray()

    # Display data
    print('data:\n', images)

if __name__ == "__main__":
    main()

The following is the output of Mult operator:

data:
[[[[6.62954025e+01 2.29007854e+01 6.44151917e+01]
  [3.36826973e+01 7.10452347e+01 8.23742752e+01]
  [5.00930786e+01 4.28848877e+01 4.84426270e+01]

  ...

  [2.04287510e+01 7.26719761e+00 9.49951267e+00]
  [3.55139351e+01 3.16886635e+01 4.68619633e+00]
  [7.50539303e+00 1.52365170e+01 1.89718781e+01]]]]