Download smoothed tensorboard values

Question

Is it possible to downloaded the smooth values generated from Tensorboard or at least get the smoothing function to be able to generate the same graphics as in Tensorboard.

did you see this? https://stackoverflow.com/questions/42011419/is-it-possible-to-call-tensorboard-smooth-function-manually — Charlie Parker, Sep 27 '21 at 20:45
Does this answer your question? [What is the mathematics behind the "smoothing" parameter in TensorBoard's scalar graphs?](https://stackoverflow.com/questions/42281844/what-is-the-mathematics-behind-the-smoothing-parameter-in-tensorboards-scalar) — Charlie Parker, Sep 27 '21 at 20:51

score 1 · Answer 1 · answered Aug 14 '17 at 19:41

1

It has changed a bit recently, but currently TensorBoard does exponential averaging for its smoothing. Should be quite easy to re-implement.

answered Aug 14 '17 at 19:41

Allen Lavoie

5,778
1
17
26

Thank you. if you know an implementation let me know – Nicolas Bruno Aug 16 '17 at 06:40

score 0 · Answer 2 · answered May 08 '20 at 14:48

0

rate = 0.1 # range of 0.0 for no smoothing, 1.0 for overly perfect smoothing.

def ema(old:float, new:float, rate:float)->float: return old * rate + new * (1.0 - rate)

ema(10, 9, 0.1)

answered May 08 '20 at 14:48

Yaoshiang

1,713
5
15

score 0 · Answer 3 · answered Sep 27 '21 at 20:50

they use some type of exp weighted avg. Try:

def my_tb_smooth(scalars: list[float], weight: float) -> list[float]:  # Weight between 0 and 1
    """

    ref: https://stackoverflow.com/questions/42011419/is-it-possible-to-call-tensorboard-smooth-function-manually

    :param scalars:
    :param weight:
    :return:
    """
    last = scalars[0]  # First value in the plot (first timestep)
    smoothed: list = []
    for point in scalars:
        smoothed_val = last * weight + (1 - weight) * point  # Calculate smoothed value
        smoothed.append(smoothed_val)                        # Save it
        last = smoothed_val                                  # Anchor the last smoothed value
    return smoothed

credit: Is it possible to call tensorboard smooth function manually?

Download smoothed tensorboard values

3 Answers3

Linked