NewYour coding agent can read the release notes before it upgrades.Set up the MCP server →
PyPI · #855 most downloaded on PyPI
TensorBoard lets you watch Tensors Flow
Last release 3 months ago
29 Jun 2026
Release timing varies
gaps range from 2 weeks to 12 months
Nearly every release is documented
notes for 53 of 53 stable releases
1 version withdrawn
withdrawn after publishing
10 years old
62 releases · first in 2017
Fix a Projector plugin vulnerability.
The 2.21 minor series tracks TensorFlow 2.21.
"Nearest Pixel" (prioritizing values closest to the cursor) (#7046, #7051, #7133, #7135).pkg_resources library in newer environments (migrated to standard importlib.metadata) (#7057).protobuf now requires >=6.31.1, <8.0.0 and grpcio now requires >=1.74.0, <2.0.The 2.21 minor series tracks TensorFlow 2.21.
"Nearest Pixel" (prioritizing values closest to the cursor) (#7046, #7051, #7133, #7135).pkg_resources library in newer environments (migrated to standard importlib.metadata) (#7057).protobuf now requires >=6.31.1, <8.0.0 and grpcio now requires >=1.74.0, <2.0.One column per quarter.
Added a dependency to Pillow (https://python-pillow.github.io/), to replace use of deprecated standard library imghdr (removed in Python 3.13) (#7009,…
The 2.20 minor series tracks TensorFlow 2.20.
imghdr (removed in Python 3.13) (#7009, thanks @Zamanhuseyinli)bleach library to 6.2.0 and removed dependency on six library. (#7013)The 2.19 minor series tracks TensorFlow 2.19.
The 2.19 minor series tracks TensorFlow 2.19.
This release contains only internal cleanups, with no particular new features or bug fixes.
The 2.18 minor series tracks TensorFlow 2.18.
The 2.18 minor series tracks TensorFlow 2.18.
protobuf restriction to work with versions < 5.0.0 (in addition to >= 5.0.0) (#6888)The 2.18 minor series tracks TensorFlow 2.18.
protobuf restriction to work with versions < 5.0.0 (in addition to >= 5.0.0) (#6888)Relax restriction on protobuf dependency
Restricts the protobuf dependency to be < 5.0.0, due to a recent breaking change (see #6808 for details).
The 2.17 minor series tracks TensorFlow 2.17.
tensorboard-data-server (aka “Rustboard”). (#6765, thanks @Corwinpro)__class__ attribute is accessed, which was blocking debugging with pyCharm (#6751, thanks @lostinplace)Remove dependency on tf-keras pip package.
Depend on tf-keras instead of tf-keras-nightly.
Not quite a breaking change, but to something to be aware of: TensorBoard releases generally follow TensorFlow’s releases. However, while TF 2.16 will…
The 2.16 minor series tracks TensorFlow 2.16.
google-auth, google-auth-oauthlib, and requests libraries. (#6705)tensorboard-data-server wheel into conformance with PEP 600, enabling support for the --load-fast flag in systems with GLIBC >= 2.29. (Requires installing/updating package tensorboard-data-server==0.7.2) (#6636 fixes #6578)tensorboard dev subcommand will now show an error message without sending any requests, as our hosted service at http://tensorboard.dev has been shut down.Updates restriction on protobuf dependency to avoid bad version while still allowing newer versions.(#6660 & #6746)
TensorBoard.dev is shutting down. See the FAQ at https://tensorboard.dev.
tensorboard dev upload command will fail. (#6638)The 2.15 minor series tracks TensorFlow 2.15.
The 2.15 minor series tracks TensorFlow 2.15.
Fix indirect dependency on the package six.
The 2.14 minor series tracks TensorFlow 2.14.
The 2.14 minor series tracks TensorFlow 2.14.
Several improvements to the projector plugin (thank you @alicialics)
Redirects the unsupported What-If Tool plugin to the supported Learning Interpretability Tool to avoid protobuf compatibility errors
Fixes scalar chart rendering in Safari
Bumps google-auth-oauthlib dependency restriction to be compatible with versions up to 1.0
google-auth-oauthlib dependency restriction to be compatible with versions up to 1.0 (#6246)The 2.12 minor series tracks TensorFlow 2.12.
The 2.12 minor series tracks TensorFlow 2.12.
--load_fast, aka “RustBoard”) improvements:
gcp_auth (#5939, thanks @Corwinpro).manylinux2014 platforms (#6101, thanks @adamjstewart).Fixes bug introduced in #6107, which was included in release 2.11.1.
Prevent regression in TensorBoard.dev uploader authentication by replacing deprecated OOB auth flow with limited-input device flow.
NOTE: This release has a bug related to the limited-input device auth flow, and was "yanked" in PyPi (https://pypi.org/help/#yanked). A newer patch release (2.11.2) contains these changes along with the fix for the bug.
The 2.11 minor series tracks TensorFlow 2.11.
The 2.11 minor series tracks TensorFlow 2.11.
Fix embedding projector plugin.
The 2.10 minor series tracks TensorFlow 2.10.
The 2.10 minor series tracks TensorFlow 2.10.
Force protobuf dependency < 3.20 to workaround incompatibilities with newer protobuf versions
The 2.9 minor series tracks TensorFlow 2.9.
The 2.9 minor series tracks TensorFlow 2.9.
The 2.8 minor series tracks TensorFlow 2.8.
The 2.8 minor series tracks TensorFlow 2.8.
learning_rate instead of lr (#5363)The 2.7 minor series tracks TensorFlow 2.7.
The 2.7 minor series tracks TensorFlow 2.7.
The 2.6 minor series tracks TensorFlow 2.6.
The 2.6 minor series tracks TensorFlow 2.6.
scalars smoothing, tooltip sorting method, ignore outliers in chart scalingreload data, reload period, pagination limitfit button in scalar chart (#4856)_output_shapes attr (#4867)tensorboard dev list (#4912)--one_shot mode (#4909)?fastChart=false fallback to old chart renderer.New data loading mode: typically loads between 100× and 350× faster
--load_fast false to disable?fastChart=false to fallback to the older version. Please give us feedback before you use the fallback mechanism.--logdir_spec (#4494) (may not have worked before)--reload_interval, which instructed TensorBoard to not read any data at allRemoved frame-ancestors * CSP directive for compatibility with Electron embeds (#4332) - thanks @joyceerhl
--path_prefix handling (#4423)frame-ancestors * CSP directive for compatibility with Electron
embeds (#4332) - thanks @joyceerhlThe 2.4 minor series tracks TensorFlow 2.4.
The 2.4 minor series tracks TensorFlow 2.4.
tensorboard dev upload on a logdir containing hparams!TensorBoard features that depend on TensorFlow APIs now require TensorFlow 2.x installed. Running TensorBoard 2.4.0+ with TensorFlow 1.x installed will not be supported.
Please note that this does not affect data already written to disk; summaries emitted by the TF 1.x tf.summary API are still readable and fully supported by the latest versions of TensorBoard.
Support for Python 3.5 is dropped.
Plugin dashboards have been removed:
Fixed deprecated numpy usage (#3768) - thanks @lgeiger
The 2.3 minor series tracks TensorFlow 2.3.
--dry_run: causes the uploader to only read the logdir and display
statistics (if --verbose is the default 1) without uploading any data to
the server--one_shot: causes the uploader to exit immediately after all existing
data in the logdir are uploaded; this mode prints a warning message if the
logdir doesn't contain any uploadable dataexport TENSORBOARD_PROXY_URL="/proxy/%PORT%/"The frontend now uses Angular (replaces the Polymer entry point, which will be
removed in a future release; still visible at the /legacy.html endpoint)
(#3779). If you observe any bugs that do not reproduce under /legacy.html,
please file an issue.
For dynamic plugins, please see their respective pages (Fairness Indicators, Profiler, What-If Tool).
Some performance improvements to line charts
tensorboard dev list subcommand now reports the total size of stored
tensors (used as the backing storage type for Histograms) (#3652)tensorboard dev upload subcommand now sends the histograms, when
available, so that it can be rendered via the Histograms plugin on
TensorBoard.dev--db_import and --db
flags) is removed to ease maintenance (#3539)TensorBoard.dev now renders model graphs, for experiments uploaded starting from this release.
tensorboard dev upload subcommand now sends the model graph, when
available, so that it can be rendered via the Graphs plugin on
TensorBoard.dev.tensorboard dev upload command supports a --plugins option to
explicitly indicate the desired plugins for which summary data should be
uploaded (#3402, #3492)tensorboard dev list subcommand now reports the total size of stored
binary objects (e.g., graphs) for each experiment (#3464)tensorboard dev list subcommand now accepts a --json flag to allow
parsing the output more easily (#3480)The 2.2 minor series tracks TensorFlow 2.2.
The 2.2 minor series tracks TensorFlow 2.2.
tensorboard-plugin-profile.
The new version works in Chrome 80 and Firefox, has better model insights and
will be more actively maintained.tensorboard dev update-metadata command allows for updating the name and
description of experiments (#3277)Uploader: Added ability to upload and modify experiment name and description
The 2.1 minor series tracks TensorFlow 2.1.
The 2.1 minor series tracks TensorFlow 2.1.
tensorboard dev list subcommand now provides detailed metadata about
each experiment.In the event of a transient backend issue or permanent breaking change, the uploader can now gracefully degrade and print a diagnostic
tensorboard dev list subcommand lists all experiments uploaded to
TensorBoard.dev (#2903)Preview of TensorBoard.dev uploader! Check out for information and usage instructions.
The 2.0 minor series tracks TensorFlow 2.0.
The 2.0 minor series tracks TensorFlow 2.0.
TensorBoard now serves on localhost only by default to avoid unintentional
overexposure. To expose TensorBoard to the network, either use a proxy, bind
to a specific hostname or IP address by using the --host flag, or explicitly
enable the previous behavior of binding on all network interfaces by passing
the flag --bind_all. See PR #2589.
The --logdir flag no longer supports passing multiple comma-delimited paths,
which means that it now supports paths containing literal comma and colon
characters, like ./logs/m=10,n=20,lr=0.001 or ./logs/run_12:30:15. To
mimic the old behavior, prefer using a tree of symlinks as it works with more
plugins, but as a fallback the flag --logdir_spec exposes the old behavior.
See PR #2664.
Projector plugin visualize_embeddings() API now takes logdir as its first
parameter rather than writer (which only supported TF 1.x summary writers).
For backwards compatibility TF 1.x writers will still be accepted, but passing
the logdir explicitly is preferred since it works without any dependency on
TF 1.x or 2.x summary writing. See PR #2665.
The namespace tensorboard.summary.* now aliases the summary API symbols in
tensorboard.summary.v2.* rather than those in tensorboard.summary.v1.*.
The old symbols can still be accessed under the .v1 names. Note that the
new v2 API symbols are exposed in TF 2.0 as the new tf.summary.* API and
this is normally how they should be used. See PR #2670.
Smarter log directory polling can be used by passing --reload_multifile=true
to poll all "active" event files in a directory rather than only the last one.
This avoids problems where data written to the non-last file never appears.
See PR #1867 for details, including how to adjust the "active" threshold.
What-If Tool now can sort PD plots by interestingness (#2461)
The 1.15 minor series tracks TensorFlow 1.15.
The 1.15 minor series tracks TensorFlow 1.15.
TENSORBOARD_BINARY environment variable can now be set to invoke a
non-default tensorboard binary (#2386)%tensorboard magic no longer spams log messages when a different
version of TensorBoard is already running on the same machine (#2470)%tensorboard magic can now be used in Jupyter notebooks running on
hosts other than localhost (#2407)summary_v2 module (#2443)tensorboard_plugins entry pointNew hyperparameters dashboard: see [tutorial and demo][hparams-docs] and [summary APIs][hparams-apis]
%load_ext tensorboard followed by %tensorboard --logdir ./path/to/logsAlt (#2221)tensorboard now supports a --version command line argument (#2097) -
thanks @shashvatshahi1998tensorboard.__version__ in addition to
tensorboard.version.VERSION (#2026).tensorboard-info is now world-writable for multi-user *nix systems
(PR #2131)importlib.reload(tensorboard) now works properly (PR #2005)## Bug fixes - #1895 - Fix strftime-related launch error on Windows (PR #1900) - #1794 - Fix What-If Tool loading examples without inference (PR #1898
strftime-related launch error on Windows (PR #1900)The 1.13 minor series tracks TensorFlow 1.13.
The 1.13 minor series tracks TensorFlow 1.13.
Compatibility note: As of 1.13, TensorBoard has begun transitioning its own use of some TensorFlow APIs to align with their naming in TF 2.0, and as a result TensorBoard 1.13+ strictly requires TensorFlow 1.13+.
%tensorboard magic, after loading the tb.notebook extension (#1813, #1822)The 1.12.2 minor series tracks TensorFlow 1.12.
The 1.12.2 minor series tracks TensorFlow 1.12.
The 1.12.1 minor series tracks TensorFlow 1.12.
The 1.12.1 minor series tracks TensorFlow 1.12.
The 1.12 minor series tracks TensorFlow 1.12.
The 1.12 minor series tracks TensorFlow 1.12.
The 1.11 minor series tracks TensorFlow 1.11.
The 1.11 minor series tracks TensorFlow 1.11.
The 1.10 minor series tracks TensorFlow 1.10.
The 1.10 minor series tracks TensorFlow 1.10.
tb.program.launch() API to launch TB from within PythonThe 1.9 minor series tracks TensorFlow 1.9.
The 1.9 minor series tracks TensorFlow 1.9.
The 1.8 minor series tracks TensorFlow 1.8.
The 1.8 minor series tracks TensorFlow 1.8.
The 1.7 minor series tracks TensorFlow 1.7.
The 1.7 minor series tracks TensorFlow 1.7.
NOTICE: TensorBoard 1.6.0+ has moved to the tensorboard package name on PyPI: https://pypi.python.org/pypi/tensorboard. Only bugfix updates on 1.5.x w
NOTICE: TensorBoard 1.6.0+ has moved to the tensorboard package name on PyPI:
https://pypi.python.org/pypi/tensorboard. Only bugfix updates on 1.5.x will be
applied to the old package name (tensorflow-tensorboard). To upgrade to
TensorBoard 1.6.0+ we suggest you first pip uninstall tensorflow-tensorboard
before doing pip install tensorboard. See "Known Issues" below if you run into
problems using TensorBoard after upgrading.
The 1.6 minor series tracks TensorFlow 1.6.
tfdbg,
the TensorFlow debugger. See for details:
https://github.com/tensorflow/tensorboard/tree/1.6/tensorboard/plugins/debuggertensorboard is installed but command and/or module are missing or
have the wrong version - this may be due to conflicts with other packages that
provide tensorboard scripts or modules. Please uninstall all such packages
including tensorboard, tensorflow-tensorboard and tb-nightly and then
reinstall tensorboard.NOTICE: TensorBoard 1.6.0+ has moved to the tensorboard package name on PyPI: https://pypi.python.org/pypi/tensorboard. Only bugfix updates on 1.5.x w
NOTICE: TensorBoard 1.6.0+ has moved to the tensorboard package name on PyPI:
https://pypi.python.org/pypi/tensorboard. Only bugfix updates on 1.5.x will be
applied to the old package name (tensorflow-tensorboard). To upgrade to TensorBoard 1.6.0+
please first pip uninstall tensorflow-tensorboard before doing pip install tensorboard.
The 1.6 minor series tracks TensorFlow 1.6. Full release notes upcoming.
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Your coding agent can read these notes before it upgrades. Set up the MCP server →