If you want to dump histograms in tensorboard while using pytorch, following [this tutorial](https://nbviewer.jupyter.org/gist/ModarTensai/b081dcf6c87f9134f29abe3a77be1ab5), you can use this.

histogram_summary.py

import torch
import tensorflow as tf


def histogram_summary(name, array):
    if not hasattr(histogram_summary, 'session'):
        histogram_summary.placeholder = tf.placeholder(tf.float32)
        histogram_summary.session = tf.Session()
        histogram_summary.histograms = {}
    if name not in histogram_summary.histograms:
        histogram_summary.histograms[name] = tf.summary.histogram(
            name, histogram_summary.placeholder)
    histogram = histogram_summary.session.run(
        histogram_summary.histograms[name],
        feed_dict={histogram_summary.placeholder: array}
    )
    return histogram


writer = tf.summary.FileWriter("exps/histogram_example")
writer.add_summary(histogram_summary('model_1/loss', torch.randn(100).numpy()))
writer.flush()