From 3f4fc5016334255d6908b20202267ca0b0287335 Mon Sep 17 00:00:00 2001
From: Jerry Ma <noreplyspamblackhole@gmail.com>
Date: Mon, 23 Sep 2019 12:25:44 -0700
Subject: [PATCH] Miscellaneous documentation improvements: (#868)

Summary:
- More clearly document the correspondence between FairseqAdam and torch.optim.AdamW
- Add ResamplingDataset to Sphinx docs
Pull Request resolved: https://github.com/fairinternal/fairseq-py/pull/868

Differential Revision: D17523244

Pulled By: jma127

fbshipit-source-id: 8e7b34b24889b2c8f70b09a52a625d2af135734b
---
 docs/data.rst         | 2 ++
 fairseq/optim/adam.py | 6 ++++++
 2 files changed, 8 insertions(+)

diff --git a/docs/data.rst b/docs/data.rst
index a2a464ec..6a390cb3 100644
--- a/docs/data.rst
+++ b/docs/data.rst
@@ -30,6 +30,8 @@ provide additional functionality:
     :members:
 .. autoclass:: fairseq.data.ConcatDataset
     :members:
+.. autoclass:: fairseq.data.ResamplingDataset
+    :members:
 .. autoclass:: fairseq.data.RoundRobinZipDatasets
     :members:
 .. autoclass:: fairseq.data.TransformEosDataset
diff --git a/fairseq/optim/adam.py b/fairseq/optim/adam.py
index 80de7f00..e60a7db5 100644
--- a/fairseq/optim/adam.py
+++ b/fairseq/optim/adam.py
@@ -15,6 +15,12 @@ from . import FairseqOptimizer, register_optimizer
 
 @register_optimizer('adam')
 class FairseqAdam(FairseqOptimizer):
+    """Adam optimizer for fairseq.
+
+    Important note: this optimizer corresponds to the "AdamW" variant of
+    Adam in its weight decay behavior. As such, it is most closely
+    analogous to torch.optim.AdamW from PyTorch.
+    """
 
     def __init__(self, args, params):
         super().__init__(args)