readme update

CHANGELOG.md

@@ -4,11 +4,29 @@ All notable changes to this project will be documented in this file.
 
 The format is based on [Keep a Changelog](http://keepachangelog.com/en/1.0.0/).
 
-
 ## [YY.MM.R] - YYYY-MM-DD
 
 ### Added
 
+-
+
+### Changed
+
+-
+
+### Removed
+
+-
+
+### Fixed
+
+-
+
+
+## [21.09] - 2021-09-30
+
+### Added
+
 - CometLogger support ([#1283](https://github.com/catalyst-team/catalyst/pull/1283))
 - CometLogger examples ([#1287](https://github.com/catalyst-team/catalyst/pull/1287))
 - XLA docs ([#1288](https://github.com/catalyst-team/catalyst/pull/1288))
@@ -37,6 +55,7 @@ The format is based on [Keep a Changelog](http://keepachangelog.com/en/1.0.0/).
 - CI/CD ([#1292](https://github.com/catalyst-team/catalyst/pull/1292)), ([#1299](https://github.com/catalyst-team/catalyst/pull/1299)), ([#1304](https://github.com/catalyst-team/catalyst/pull/1304)), ([#1306](https://github.com/catalyst-team/catalyst/pull/1306))
 - Optuna configs ([#1296](https://github.com/catalyst-team/catalyst/pull/1292)), ([#1296](https://github.com/catalyst-team/catalyst/pull/1299))
 
+
 ## [21.08] - 2021-08-31
 
 ### Added

catalyst/__version__.py

@@ -1 +1 @@
-__version__ = "21.09"
+__version__ = "21.10rc0"

catalyst/callbacks/metrics/r2_squared.py

@@ -47,7 +47,7 @@ class R2SquaredCallback(LoaderMetricCallback):
             num_epochs=8,
             verbose=True,
             callbacks=[
-                      dl.R2SquaredCallback(input_key="logits", target_key="targets")
+                dl.R2SquaredCallback(input_key="logits", target_key="targets")
             ]
         )
 

docs/api/callbacks.rst

@@ -275,6 +275,13 @@ RecSys – NDCGCallback
     :exclude-members: __init__, on_experiment_start, on_stage_start, on_epoch_start, on_loader_start, on_batch_start, on_batch_end, on_loader_end, on_epoch_end, on_stage_end, on_experiment_end
     :show-inheritance:
 
+Regression – R2SquaredCallback
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+.. autoclass:: catalyst.callbacks.metrics.r2_squared.R2SquaredCallback
+    :members:
+    :exclude-members: __init__, on_experiment_start, on_stage_start, on_epoch_start, on_loader_start, on_batch_start, on_batch_end, on_loader_end, on_epoch_end, on_stage_end, on_experiment_end
+    :show-inheritance:
+
 Segmentation – DiceCallback
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~
 .. autoclass:: catalyst.callbacks.metrics.segmentation.DiceCallback

docs/api/contrib.rst

@@ -153,6 +153,20 @@ MarginLoss
     :undoc-members:
     :show-inheritance:
 
+NTXentLoss
+""""""""""
+.. autoclass:: catalyst.contrib.nn.criterion.ntxent.NTXentLoss
+    :members: __init__
+    :undoc-members:
+    :show-inheritance:
+
+SupervisedContrastiveLoss
+"""""""""""""""""""""""""
+.. autoclass:: catalyst.contrib.nn.criterion.supervised_contrastive.SupervisedContrastiveLoss
+    :members: __init__
+    :undoc-members:
+    :show-inheritance:
+
 TrevskyLoss
 """""""""""
 .. autoclass:: catalyst.contrib.nn.criterion.trevsky.TrevskyLoss
@@ -181,34 +195,34 @@ WingLoss
     :undoc-members:
     :show-inheritance:
 
-Regression
-""""""""""
+Contrastive
+"""""""""""
 
-HuberLossV0
-*****************
-.. autoclass:: catalyst.contrib.nn.criterion.regression.HuberLossV0
-    :members:
+BarlowTwinsLoss
+***************
+.. autoclass:: catalyst.contrib.nn.criterion.contrastive.BarlowTwinsLoss
+    :members: __init__
     :undoc-members:
     :show-inheritance:
 
-CategoricalRegressionLoss
-*************************
-.. autoclass:: catalyst.contrib.nn.criterion.regression.CategoricalRegressionLoss
-    :members:
+ContrastiveDistanceLoss
+***********************
+.. autoclass:: catalyst.contrib.nn.criterion.contrastive.ContrastiveDistanceLoss
+    :members: __init__
     :undoc-members:
     :show-inheritance:
 
-QuantileRegressionLoss
-*************************
-.. autoclass:: catalyst.contrib.nn.criterion.regression.QuantileRegressionLoss
-    :members:
+ContrastiveEmbeddingLoss
+************************
+.. autoclass:: catalyst.contrib.nn.criterion.contrastive.ContrastiveEmbeddingLoss
+    :members: __init__
     :undoc-members:
     :show-inheritance:
 
-RSquareLoss
-*************************
-.. autoclass:: catalyst.contrib.nn.criterion.regression.RSquareLoss
-    :members:
+ContrastivePairwiseEmbeddingLoss
+********************************
+.. autoclass:: catalyst.contrib.nn.criterion.contrastive.ContrastivePairwiseEmbeddingLoss
+    :members: __init__
     :undoc-members:
     :show-inheritance:
 
@@ -257,6 +271,36 @@ WARPLoss
     :undoc-members:
     :show-inheritance:
 
+Regression
+""""""""""
+
+HuberLossV0
+*****************
+.. autoclass:: catalyst.contrib.nn.criterion.regression.HuberLossV0
+    :members:
+    :undoc-members:
+    :show-inheritance:
+
+CategoricalRegressionLoss
+*************************
+.. autoclass:: catalyst.contrib.nn.criterion.regression.CategoricalRegressionLoss
+    :members:
+    :undoc-members:
+    :show-inheritance:
+
+QuantileRegressionLoss
+*************************
+.. autoclass:: catalyst.contrib.nn.criterion.regression.QuantileRegressionLoss
+    :members:
+    :undoc-members:
+    :show-inheritance:
+
+RSquareLoss
+*************************
+.. autoclass:: catalyst.contrib.nn.criterion.regression.RSquareLoss
+    :members:
+    :undoc-members:
+    :show-inheritance:
 
 Modules
 ~~~~~~~~~~~~~~~~

docs/faq/data.rst

@@ -232,7 +232,7 @@ Suppose, you are using set of different metrics in your pipeline:
         def handle_batch(self, batch):
             # on this step we also have self.batch = batch = {"features": ..., "targets": ...}
             logits = self.model(batch["features"])
-            self.output = {"logits": logits}
+            self.batch = {**{"logits": logits}, **self.batch}
             # this is useful for other components of the pipeline
 
     loaders = {"train": ..., "valid": ...}
@@ -261,7 +261,7 @@ For example, to use ``accuracy01`` metric as your model selection one you could
         def handle_batch(self, batch):
             # on this step we also have self.batch = batch = {"features": ..., "targets": ...}
             logits = self.model(batch["features"])
-            self.output = {"logits": logits}
+            self.batch = {**{"logits": logits}, **self.batch}
             # this is useful for other components of the pipeline
 
     loaders = {"train": ..., "valid": ...}

docs/index.rst

@@ -99,20 +99,17 @@ Getting started
 
 Step by step guide
 ~~~~~~~~~~~~~~~~~~~~~~
-1. Start with `Catalyst 2021–Accelerated PyTorch 2.0`_ introduction.
-2. Check `minimal examples`_.
-3. Try `notebook tutorials with Google Colab`_.
-4. Read `blogposts`_ with use-cases and guides.
-5. Learn machine learning with our `"Deep Learning with Catalyst" course`_.
-6. If you would like to contribute to the project, follow our `contribution guidelines`_.
-7. If you want to support the project, feel free to donate on `patreon page`_ or `write us`_ with your proposals.
-8. And do not forget to `join our slack`_ for collaboration.
-
-.. _`Catalyst 2021–Accelerated PyTorch 2.0`: https://medium.com/catalyst-team/catalyst-2021-accelerated-pytorch-2-0-850e9b575cb6?source=friends_link&sk=865d3c472cfb10379864656fedcfe762
+1. Start with `Catalyst — A PyTorch Framework for Accelerated Deep Learning R&D`_ introduction.
+2. Try `notebook tutorials`_ or check `minimal examples`_ for first deep dive.
+3. Read `blogposts`_ with use-cases and guides.
+4. Learn machine learning with our `"Deep Learning with Catalyst" course`_.
+5. And do not forget to `join our slack`_ for collaboration.
+
+.. _`Catalyst — A PyTorch Framework for Accelerated Deep Learning R&D`: https://medium.com/pytorch/catalyst-a-pytorch-framework-for-accelerated-deep-learning-r-d-ad9621e4ca88?source=friends_link&sk=885b4409aecab505db0a63b06f19dcef
 .. _`Kittylyst`: https://github.com/Scitator/kittylyst
 .. _`minimal examples`: https://github.com/catalyst-team/catalyst#minimal-examples
-.. _`Notebook Tutorials with Google Colab`: https://github.com/catalyst-team/catalyst#tutorials
-.. _`blogposts`: https://github.com/catalyst-team/catalyst#blogposts
+.. _`notebook tutorials`: https://github.com/catalyst-team/catalyst#minimal-examples
+.. _`blogposts`: https://catalyst-team.com/post/
 .. _`"Deep Learning with Catalyst" course`: https://github.com/catalyst-team/dl-course
 .. _`classification`: https://github.com/catalyst-team/classification
 .. _`detection`: https://github.com/catalyst-team/detection