Commits

Branch selector

 master

User selector

mdreves

Datepicker

All time

Commit History

Commits on Dec 15, 2021

• Convert all failed label lookups to one-hot vectors of all 0's when possible so that confusion matrix metrics that rely label using 0 or 1 still work.

  Show description for 2a61e5a
  

  mdreves

  authored and

  tfx-copybara

  committed
  2a61e5a

  Copy full SHA for 2a61e5a

Commits on Nov 8, 2021

• Automated rollback of commit 65ac581

  Show description for 36d60e0
  

  mdreves

  authored and

  tfx-copybara

  committed
  36d60e0

  Copy full SHA for 36d60e0

Commits on Nov 4, 2021

• Part I of replacing keras metric types with TFMA versions: confusion matrix metrics.

  Show description for 67b71d2
  

  mdreves

  authored and

  tfx-copybara

  committed
  67b71d2

  Copy full SHA for 67b71d2

Commits on Nov 3, 2021

• Update query_statistics to support weighted examples.

  Show description for 4180df6
  

  mdreves

  authored and

  tfx-copybara

  committed
  4180df6

  Copy full SHA for 4180df6

Commits on Oct 27, 2021

• Add support for specifying if a metric is weighted or not.

  Show description for 9c90d06
  

  mdreves

  authored and

  tfx-copybara

  committed
  9c90d06

  Copy full SHA for 9c90d06

Commits on Oct 26, 2021

• Add support for checking np.floating as well as float when summing attributions.

  Show description for 80a7dfb
  

  mdreves

  authored and

  tfx-copybara

  committed
  80a7dfb

  Copy full SHA for 80a7dfb

Commits on Oct 20, 2021

• noop

  Show description for c8fc444
  

  mdreves

  authored and

  tfx-copybara

  committed
  c8fc444

  Copy full SHA for c8fc444

Commits on Oct 7, 2021

• Update JS files to recent change for JupyterLab fix.

  Show description for 50fe33a
  

  mdreves

  authored and

  tfx-copybara

  committed
  50fe33a

  Copy full SHA for 50fe33a

Commits on Oct 6, 2021

• Add support for non-scalar example weights.

  Show description for d9ca830
  

  mdreves

  authored and

  tfx-copybara

  committed
  d9ca830

  Copy full SHA for d9ca830

Commits on Sep 16, 2021

• Move top-level classes into utils subdirectory to avoid python package name collisions.

  Show description for 0e79174
  

  mdreves

  authored and

  tf-model-analysis-team

  committed
  0e79174

  Copy full SHA for 0e79174

Commits on Sep 2, 2021

• Fix invalid error message formatting.

  Show description for d2dbd2f
  

  mdreves

  authored and

  tf-model-analysis-team

  committed
  d2dbd2f

  Copy full SHA for d2dbd2f

Commits on Aug 26, 2021

• Fixed bug where example_count post_export_metric is added even if include_default_metrics is False.

  Show description for 557a007
  

  mdreves

  authored and

  tf-model-analysis-team

  committed
  557a007

  Copy full SHA for 557a007

Commits on Aug 25, 2021

• noop

  Show description for 291eeaa
  

  mdreves

  authored and

  tf-model-analysis-team

  committed
  291eeaa

  Copy full SHA for 291eeaa

Commits on Aug 21, 2021

• noop

  Show description for 206514b
  

  mdreves

  authored and

  tf-model-analysis-team

  committed
  206514b

  Copy full SHA for 206514b

• Add better error messages for predictions, labels, and example weights that are used with multi-output models.

  Show description for 7aceefb
  

  mdreves

  authored and

  tf-model-analysis-team

  committed
  7aceefb

  Copy full SHA for 7aceefb

Commits on Aug 17, 2021

• Added SparseTensorValue and RaggedTensorValue types for storing in-memory versions of sparse and ragged tensor values used in extracts. Tensor values used for features, etc should now be based on e…

  Show description for 606872d
  

  mdreves

  authored and

  tf-model-analysis-team

  committed
  606872d

  Copy full SHA for 606872d

Commits on Aug 13, 2021

• Add FAQ on the need for binarization settings when a binary classification model outputs two classes.

  Show description for d2a416c
  

  mdreves

  authored and

  tf-model-analysis-team

  committed
  d2a416c

  Copy full SHA for d2a416c

• Update to use importlib for imports used for registration purposes only.

  Show description for e879b7a
  

  mdreves

  authored and

  tf-model-analysis-team

  committed
  e879b7a

  Copy full SHA for e879b7a

Commits on Aug 12, 2021

• Updated comments for metric_specs.output_names to indicate that it is required for multi-output models.

  Show description for 380640d
  

  mdreves

  authored and

  tf-model-analysis-team

  committed
  380640d

  Copy full SHA for 380640d

Commits on Aug 11, 2021

• noop

  Show description for 7487680
  

  mdreves

  authored and

  tf-model-analysis-team

  committed
  7487680

  Copy full SHA for 7487680

Commits on Aug 3, 2021

• Handle cases where model output names is None.

  Show description for 3a0a5d6
  

  mdreves

  authored and

  tf-model-analysis-team

  committed
  3a0a5d6

  Copy full SHA for 3a0a5d6

• Update comment to indicate that 'scores' is an TF estimator output key that will automatically be searched for along with probabilities, etc.

  Show description for 09fa445
  

  mdreves

  authored and

  tf-model-analysis-team

  committed
  09fa445

  Copy full SHA for 09fa445

Commits on Jul 27, 2021

• Update comment for config field within MetricConfig to specify that it applies to the from_config method not the __init__ method.

  Show description for 384a5cf
  

  mdreves

  authored and

  tf-model-analysis-team

  committed
  384a5cf

  Copy full SHA for 384a5cf

Commits on Jul 8, 2021

• Fix issue with string encodings of float values being disallowed.

  Show description for 31b9f39
  

  mdreves

  authored and

  tf-model-analysis-team

  committed
  31b9f39

  Copy full SHA for 31b9f39

Commits on Apr 2, 2021

• Move some classes to private that were not intended to be public.

  Show description for 2cee2dc
  

  mdreves

  authored and

  tf-model-analysis-team

  committed
  2cee2dc

  Copy full SHA for 2cee2dc

Commits on Mar 30, 2021

• Add support for keras model.compiled_metrics and model.compiled_loss that are None.

  Show description for 78a01f6
  

  mdreves

  authored and

  tf-model-analysis-team

  committed
  78a01f6

  Copy full SHA for 78a01f6

Commits on Mar 22, 2021

• Add support for empty input features (i.e. all features added via preprocessing functions).

  Show description for 191d486
  

  mdreves

  authored and

  tf-model-analysis-team

  committed
  191d486

  Copy full SHA for 191d486

• Add support for output aggregation.

  Show description for e4c6f89
  

  mdreves

  authored and

  tf-model-analysis-team

  committed
  e4c6f89

  Copy full SHA for e4c6f89

Commits on Mar 10, 2021

• Audited Combiner based implementations to ensure using Beam APIs efficiently.

  Show description for 63ae2cb
  

  mdreves

  authored and

  tf-model-analysis-team

  committed
  63ae2cb

  Copy full SHA for 63ae2cb

• Fix invalid calculations for metrics derived from tf.keras.losses.

  Show description for a24d82d
  

  mdreves

  authored and

  tf-model-analysis-team

  committed
  a24d82d

  Copy full SHA for a24d82d

Commits on Mar 9, 2021

• Fix failing tests that were not compatible with TensorFlow==1.15.

  Show description for fc29651
  

  mdreves

  authored and

  tf-model-analysis-team

  committed
  fc29651

  Copy full SHA for fc29651

Commits on Mar 6, 2021

• Updated TODO

  Show description for 392f684
  

  mdreves

  authored and

  tf-model-analysis-team

  committed
  392f684

  Copy full SHA for 392f684

Commits on Mar 4, 2021

• Add PerSliceMetricThresholds to __init__.py

  Show description for bcdf72f
  

  mdreves

  authored and

  tf-model-analysis-team

  committed
  bcdf72f

  Copy full SHA for bcdf72f

Commits on Mar 3, 2021

• noop

  Show description for 4f6c35f
  

  mdreves

  authored and

  tf-model-analysis-team

  committed
  4f6c35f

  Copy full SHA for 4f6c35f

• Remove pytype conversions.

  Show description for ff03c62
  

  mdreves

  authored and

  tf-model-analysis-team

  committed
  ff03c62

  Copy full SHA for ff03c62

Pagination

Previous

Next