tqdm

Py-Versions Versions Conda-Forge-Status Docker Snapcraft
Build-Status Coverage-Status Branch-Coverage-Status Codacy-Grade Libraries-Rank PyPI-Downloads
DOI LICENCE OpenHub-Status binder-demo awesome-python

tqdm derives from the Arabic word taqaddum (تقدّم) which can mean "progress," and is an abbreviation for "I love you so much" in Spanish (te quiero demasiado).

Instantly make your loops show a smart progress meter - just wrap any iterable with tqdm(iterable), and you're done!

In [ ]:
from tqdm import tqdm
for i in tqdm(range(10000)):
    pass

trange(N) can be also used as a convenient shortcut for tqdm(xrange(N)).

In [ ]:
from tqdm import trange
for i in trange(10000, unit_scale=True, desc="hello", unit="epoch"):
    pass
Screenshot Video Slides Merch

It can also be executed as a module with pipes:

In [ ]:
! seq 9999999 | tqdm --bytes | wc -l
tar -zcf - docs/ | tqdm --bytes --total `du -sb docs/ | cut -f1` > backup.tgz
 44%|██████████████▊                   | 153M/352M [00:14<00:18, 11.0MB/s]

Overhead is low -- about 60ns per iteration (80ns with tqdm.gui), and is unit tested against performance regression. By comparison, the well-established ProgressBar has an 800ns/iter overhead.

In addition to its low overhead, tqdm uses smart algorithms to predict the remaining time and to skip unnecessary iteration displays, which allows for a negligible overhead in most cases.

tqdm works on any platform (Linux, Windows, Mac, FreeBSD, NetBSD, Solaris/SunOS), in any console or in a GUI, and is also friendly with IPython/Jupyter notebooks.

tqdm does not require any dependencies (not even curses!), just Python and an environment supporting carriage return \r and line feed \n control characters.


Usage

tqdm is very versatile and can be used in a number of ways. The three main ones are given below.

Iterable-based

Wrap tqdm() around any iterable:

In [ ]:
from tqdm import tqdm
from time import sleep
In [ ]:
text = ""
for char in tqdm(["a", "b", "c", "d"]):
    sleep(0.25)
    text = text + char

trange(i) is a special optimised instance of tqdm(range(i)):

In [ ]:
from tqdm import trange

for i in trange(100):
    sleep(0.01)

Instantiation outside of the loop allows for manual control over tqdm():

In [ ]:
pbar = tqdm(["a", "b", "c", "d"])
for char in pbar:
    sleep(0.25)
    pbar.set_description("Processing %s" % char)

Manual

Manual control of tqdm() updates using a with statement:

In [ ]:
with tqdm(total=100) as pbar:
    for i in range(10):
        sleep(0.1)
        pbar.update(10)

If the optional variable total (or an iterable with len()) is provided, predictive stats are displayed.

with is also optional (you can just assign tqdm() to a variable, but in this case don't forget to del or close() at the end:

In [ ]:
pbar = tqdm(total=100)
for i in range(10):
    sleep(0.1)
    pbar.update(10)
pbar.close()

Module

Perhaps the most wonderful use of tqdm is in a script or on the command line. Simply inserting tqdm (or python -m tqdm) between pipes will pass through all stdin to stdout while printing progress to stderr.

The example below demonstrated counting the number of lines in all Python files in the current directory, with timing information included.

In [ ]:
! time find . -name '*.py' -type f -exec cat \{} \; | wc -l
In [ ]:
! time find . -name '*.py' -type f -exec cat \{} \; | tqdm | wc -l

Note that the usual arguments for tqdm can also be specified.

In [ ]:
! find . -name '*.py' -type f -exec cat \{} \; | tqdm --unit loc --unit-scale --total 4104300 --null

Backing up a large directory?

tar -zcf - docs/ | tqdm --bytes --total `du -sb docs/ | cut -f1` > backup.tgz
 44%|██████████████▊                   | 153M/352M [00:14<00:18, 11.0MB/s]

This can be beautified further:

BYTES="$(du -sb docs/ | cut -f1)"
tar -cf - docs/ \
  | tqdm --bytes --total "$BYTES" --desc Processing | gzip \
  | tqdm --bytes --total "$BYTES" --desc Compressed --position 1 \
  > ~/backup.tgz
Processing: 100%|██████████████████████| 352M/352M [00:14<00:00, 30.2MB/s]
Compressed:  42%|█████████▎            | 148M/352M [00:14<00:19, 10.9MB/s]

Or done on a file level using 7-zip:

7z a -bd -r backup.7z docs/ | grep Compressing \
  | tqdm --total $(find docs/ -type f | wc -l) --unit files \
  | grep -v Compressing
100%|██████████████████████████▉| 15327/15327 [01:00<00:00, 712.96files/s]

Documentation

In [ ]:
tqdm?
In [ ]:
! tqdm --help

Examples and Advance Usage

Description and additional stats

Custom information can be displayed and updated dynamically on tqdm bars with the desc and postfix arguments:

In [ ]:
from tqdm import tqdm, trange
from random import random, randint
from time import sleep

with trange(10) as t:
    for i in t:
        # Description will be displayed on the left
        t.set_description('GEN %i' % i)
        # Postfix will be displayed on the right,
        # formatted automatically based on argument's datatype
        t.set_postfix(loss=random(), gen=randint(1,999), str='h',
                      lst=[1, 2])
        sleep(0.1)

with tqdm(total=10, bar_format="{postfix[0]} {postfix[1][value]:>8.2g}",
          postfix=["Batch", dict(value=0)]) as t:
    for i in range(10):
        sleep(0.1)
        t.postfix[1]["value"] = i / 2
        t.update()

Points to remember when using {postfix[...]} in the bar_format string:

  • postfix also needs to be passed as an initial argument in a compatible format, and
  • postfix will be auto-converted to a string if it is a dict-like object. To prevent this behaviour, insert an extra item into the dictionary where the key is not a string.

Additional bar_format parameters may also be defined by overriding format_dict, and the bar itself may be modified using ascii:

In [ ]:
from tqdm import tqdm
from time import sleep

class TqdmExtraFormat(tqdm):
    """Provides a `total_time` format parameter"""
    @property
    def format_dict(self):
        d = super(TqdmExtraFormat, self).format_dict
        total_time = d["elapsed"] * (d["total"] or 0) / max(d["n"], 1)
        d.update(total_time=self.format_interval(total_time) + " in total")
        return d

for i in TqdmExtraFormat(
      range(9), ascii=" .oO0",
      bar_format="{total_time}: {percentage:.0f}%|{bar}{r_bar}"):
    if i == 4:
        break

Note that {bar} also supports a format specifier [width][type].

  • width
    • unspecified (default): automatic to fill ncols
    • int >= 0: fixed width overriding ncols logic
    • int < 0: subtract from the automatic default
  • type
    • a: ascii (ascii=True override)
    • u: unicode (ascii=False override)
    • b: blank (ascii=" " override)

This means a fixed bar with right-justified text may be created by using: bar_format="{l_bar}{bar:10}|{bar:-10b}right-justified"

Nested progress bars

tqdm supports nested progress bars. Here's an example:

In [ ]:
from tqdm.auto import trange
from time import sleep

for i in trange(4, desc='1st loop'):
    for j in trange(5, desc='2nd loop'):
        for k in trange(50, desc='3rd loop', leave=False):
            sleep(0.01)

For manual control over positioning (e.g. for multi-processing use), you may specify position=n where n=0 for the outermost bar, n=1 for the next, and so on. However, it's best to check if tqdm can work without manual position first.

from time import sleep
from tqdm import trange, tqdm
from multiprocessing import Pool, RLock, freeze_support

L = list(range(9))

def progresser(n):
    interval = 0.001 / (n + 2)
    total = 5000
    text = "#{}, est. {:<04.2}s".format(n, interval * total)
    for _ in trange(total, desc=text, position=n):
        sleep(interval)

if __name__ == '__main__':
    freeze_support()  # for Windows support
    tqdm.set_lock(RLock())  # for managing output contention
    p = Pool(initializer=tqdm.set_lock, initargs=(tqdm.get_lock(),))
    p.map(progresser, L)

Note that in Python 3, tqdm.write is thread-safe:

from time import sleep
from tqdm import tqdm, trange
from concurrent.futures import ThreadPoolExecutor

L = list(range(9))

def progresser(n):
    interval = 0.001 / (n + 2)
    total = 5000
    text = "#{}, est. {:<04.2}s".format(n, interval * total)
    for _ in trange(total, desc=text):
        sleep(interval).auto
    if n == 6:
        tqdm.write("n == 6 completed.")
        tqdm.write("`tqdm.write()` is thread-safe in py3!")

if __name__ == '__main__':
    with ThreadPoolExecutor() as p:
        p.map(progresser, L)

Hooks and callbacks

tqdm can easily support callbacks/hooks and manual updates. Here's an example with urllib:

urllib.urlretrieve documentation

[...] If present, the hook function will be called once on establishment of the network connection and once after each block read thereafter. The hook will be passed three arguments; a count of blocks transferred so far, a block size in bytes, and the total size of the file. [...]

In [ ]:
import urllib, os
from tqdm import tqdm
urllib = getattr(urllib, 'request', urllib)

class TqdmUpTo(tqdm):
    """Provides `update_to(n)` which uses `tqdm.update(delta_n)`."""
    def update_to(self, b=1, bsize=1, tsize=None):
        """
        b  : int, optional
            Number of blocks transferred so far [default: 1].
        bsize  : int, optional
            Size of each block (in tqdm units) [default: 1].
        tsize  : int, optional
            Total size (in tqdm units). If [default: None] remains unchanged.
        """
        if tsize is not None:
            self.total = tsize
        return self.update(b * bsize - self.n)  # also sets self.n = b * bsize

eg_link = "https://caspersci.uk.to/matryoshka.zip"
with TqdmUpTo(unit='B', unit_scale=True, unit_divisor=1024, miniters=1,
              desc=eg_link.split('/')[-1]) as t:  # all optional kwargs
    urllib.urlretrieve(eg_link, filename=os.devnull,
                       reporthook=t.update_to, data=None)
    t.total = t.n

Inspired by twine#242. Functional alternative in examples/tqdm_wget.py.

It is recommend to use miniters=1 whenever there is potentially large differences in iteration speed (e.g. downloading a file over a patchy connection).

Wrapping read/write methods

To measure throughput through a file-like object's read or write methods, use CallbackIOWrapper:

from tqdm import tqdm
from tqdm.utils import CallbackIOWrapper

with tqdm(total=file_obj.size,
          unit='B', unit_scale=True, unit_divisor=1024) as t:
    fobj = CallbackIOWrapper(t.update, file_obj, "read")
    while True:
        chunk = fobj.read(chunk_size)
        if not chunk:
            break
    t.reset()
    # ... continue to use `t` for something else

Alternatively, use the even simpler wrapattr convenience function, which would condense both the urllib and CallbackIOWrapper examples down to:

In [ ]:
import urllib, os
from tqdm import tqdm

eg_link = "https://caspersci.uk.to/matryoshka.zip"
response = getattr(urllib, 'request', urllib).urlopen(eg_link)
with tqdm.wrapattr(open(os.devnull, "wb"), "write",
                   miniters=1, desc=eg_link.split('/')[-1],
                   total=getattr(response, 'length', None)) as fout:
    for chunk in response:
        fout.write(chunk)

The requests equivalent is nearly identical:

In [ ]:
import requests, os
from tqdm import tqdm

eg_link = "https://caspersci.uk.to/matryoshka.zip"
response = requests.get(eg_link, stream=True)
with tqdm.wrapattr(open(os.devnull, "wb"), "write",
                   miniters=1, desc=eg_link.split('/')[-1],
                   total=int(response.headers.get('content-length', 0))) as fout:
    for chunk in response.iter_content(chunk_size=4096):
        fout.write(chunk)

Pandas Integration

Due to popular demand we've added support for pandas -- here's an example for DataFrame.progress_apply and DataFrameGroupBy.progress_apply:

In [ ]:
import pandas as pd
import numpy as np
from tqdm import tqdm

df = pd.DataFrame(np.random.randint(0, 100, (100000, 6)))

# Register `pandas.progress_apply` and `pandas.Series.map_apply` with `tqdm`
# (can use `tqdm.gui.tqdm`, `tqdm.notebook.tqdm`, optional kwargs, etc.)
tqdm.pandas(desc="my bar!")

# Now you can use `progress_apply` instead of `apply`
# and `progress_map` instead of `map`
df.progress_apply(lambda x: x**2)
# can also groupby:
# df.groupby(0).progress_apply(lambda x: x**2)

In case you're interested in how this works (and how to modify it for your own callbacks), see the examples folder or import the module and run help().

Keras Integration

A keras callback is also available:

from tqdm.keras import TqdmCallback

...

model.fit(..., verbose=0, callbacks=[TqdmCallback()])

IPython/Jupyter Integration

IPython/Jupyter is supported via the tqdm.notebook submodule:

In [ ]:
from tqdm.notebook import trange, tqdm
from time import sleep

for i in trange(3, desc='1st loop'):
    for j in tqdm(range(100), desc='2nd loop'):
        sleep(0.01)

In addition to tqdm features, the submodule provides a native Jupyter widget (compatible with IPython v1-v4 and Jupyter), fully working nested bars and colour hints (blue: normal, green: completed, red: error/interrupt, light blue: no ETA); as demonstrated below.

Screenshot-Jupyter3

The notebook version supports percentage or pixels for overall width (e.g.: ncols='100%' or ncols='480px').

It is also possible to let tqdm automatically choose between console or notebook versions by using the autonotebook submodule:

In [ ]:
from tqdm.autonotebook import tqdm
tqdm.pandas()

Note that this will issue a TqdmExperimentalWarning if run in a notebook since it is not meant to be possible to distinguish between jupyter notebook and jupyter console. Use auto instead of autonotebook to suppress this warning.

Note that notebooks will display the bar in the cell where it was created. This may be a different cell from the one where it is used. If this is not desired, the creation of the bar must be delayed/moved to the cell where it is desired to be displayed.

Another possibility is to have a single bar (near the top of the notebook) which is constantly re-used (using reset() rather than close()). For this reason, the notebook version (unlike the CLI version) does not automatically call close() upon Exception.

In [ ]:
from tqdm.notebook import tqdm
pbar = tqdm()
In [ ]:
# different cell
iterable = range(100)
pbar.reset(total=len(iterable))  # initialise with new `total`
for i in iterable:
    pbar.update()
pbar.refresh()  # force print final status but don't `close()`

Writing messages

This is a work in progress (see #737).

Since tqdm uses a simple printing mechanism to display progress bars, you should not write any message in the terminal using print() while a progressbar is open.

To write messages in the terminal without any collision with tqdm bar display, a .write() method is provided:

In [ ]:
from tqdm.auto import tqdm, trange
from time import sleep

bar = trange(10)
for i in bar:
    # Print using tqdm class method .write()
    sleep(0.1)
    if not (i % 3):
        tqdm.write("Done task %i" % i)
    # Can also use bar.write()

By default, this will print to standard output sys.stdout. but you can specify any file-like object using the file argument. For example, this can be used to redirect the messages writing to a log file or class.

README-Hits|(Since 19 May 2016) -|-

Do your own experiments here 👇

Try tqdm youself by adding your code below and running your own experiments.

In [ ]:
import tqdm

# your code here
tqdm.