Refine the format of printing tensor

[zhiqiu]

PR types

Others

PR changes

Others

Describe

Refine the format of printing tensor

• Provides a new API, paddle.set_printoptions(precision=None, threshold=None, edgeitems=None, sci_mode=None), which is similar with numpy.set_printoptions. See details in the code comments.

• Print short forms when the size of the Tensor exceeds the threshold.

For example, the following example code,

• Before

import paddle
print(paddle.rand([10,10]))

Tensor: eager_tmp_0
  - place: CPUPlace
  - shape: [10, 10]
  - layout: NCHW
  - dtype: float
  - data: [0.533863 0.0797833 0.221516 0.812535 0.886828 0.0622117 0.0192084 0.20787 0.0820772 0.948683 0.888249 0.67202 0.0943131 0.589427 0.232097 0.561007 0.39321 0.649692 0.261727 0.810542 0.671124 0.830243 0.658425 0.825302 0.129626 0.17518 0.106716 0.268798 0.135505 0.972766 0.391849 0.931962 0.716626 0.259243 0.902846 0.302401 0.208576 0.182095 0.915462 0.25003 0.21024 0.484252 0.319637 0.216295 0.798256 0.61407 0.753209 0.168424 0.988953 0.735267 0.0641483 0.501011 0.283179 0.330886 0.218066 0.260493 0.237336 0.426256 0.824428 0.965626 0.732239 0.896424 0.293316 0.669529 0.847223 0.315755 0.787443 0.646005 0.320226 0.644231 0.676818 0.659425 0.101772 0.616528 0.40214 0.719995 0.864722 0.98008 0.454678 0.753384 0.601966 0.483612 0.421961 0.482417 0.736744 0.36805 0.140307 0.781247 0.55052 0.16105 0.0330559 0.642412 0.889036 0.664003 0.089531 0.749408 0.696497 0.608919 0.148653 0.637941]

• After

import paddle

paddle.manual_seed(10)
a = paddle.rand([10, 20])
paddle.set_printoptions(4, 100, 3)
print(a)

linear = paddle.nn.Linear(3, 5)
print(linear.weight)

Tensor(shape=[10, 20], dtype=float32, place=CPUPlace, stop_gradient=True,
       [[0.2727, 0.5489, 0.8655, ..., 0.2916, 0.8525, 0.9000],
        [0.3806, 0.8996, 0.0928, ..., 0.9535, 0.8378, 0.6409],
        [0.1484, 0.4038, 0.8294, ..., 0.0148, 0.6520, 0.4250],
        ...,
        [0.3426, 0.1909, 0.7240, ..., 0.4218, 0.2676, 0.5679],
        [0.5561, 0.2081, 0.0676, ..., 0.9778, 0.3302, 0.9559],
        [0.2665, 0.8483, 0.5389, ..., 0.4956, 0.6862, 0.9178]])

Parameter containing:
Tensor(shape=[3, 5], dtype=float32, place=CPUPlace, stop_gradient=False,
       [[-0.5238,  0.1450, -0.1731, -0.1135,  0.6033],
        [ 0.5953, -0.3029,  0.8256,  0.6461, -0.5687],
        [-0.0755, -0.5802, -0.3434, -0.6548,  0.8353]])

[paddle-bot-old]

Thanks for your contribution!
Please wait for the result of CI firstly. See Paddle CI Manual for details.

[dingjiaweiww]

LGTM

[luotao1]

LGTM for framework.py

[jzhang533]

lgtm

[luotao1]

LGTM