Merge branch 'master' into master

Author: cyfra

.github/pull_request_template.md

@@ -0,0 +1,20 @@
+# Add Dataset
+
+* Dataset Name: <name>
+* Issue Reference: <link>
+* `dataset_info.json` Gist: <link>
+  
+## Description
+
+<description>
+  
+## Checklist
+* [ ] Address all TODO's
+* [ ] Add alphabetized import to subdirectory's `__init__.py`
+* [ ] Run `download_and_prepare` successfully
+* [ ] Add checksums file
+* [ ] Properly cite in `BibTeX` format
+* [ ] Add passing test(s)
+* [ ] Add test data
+* [ ] Add data generation script (if applicable)
+* [ ] Lint code

README.md

@@ -145,7 +145,7 @@ input pipelines with `tf.data` but use whatever you'd like for your model
 components.
 
 ```python
-train_ds = tfds.load("mnist", split=tfds.Split.TRAIN)
+train_ds = tfds.load("mnist", split="train")
 train_ds = train_ds.shuffle(1024).batch(128).repeat(5).prefetch(10)
 for example in tfds.as_numpy(train_ds):
   numpy_images, numpy_labels = example["image"], example["label"]

docs/_book.yaml

@@ -27,7 +27,7 @@ upper_tabs:
         path: /datasets/add_dataset
       - title: Feature decoding
         path: /datasets/decode
-      - title: Add huge datasets
+      - title: Add huge datasets (Apache Beam)
         path: /datasets/beam_datasets
       - title: Store your dataset on GCS
         path: /datasets/gcs

docs/_index.ipynb

@@ -47,7 +47,7 @@
         "print(tfds.list_builders())\n",
         "\n",
         "# Construct a tf.data.Dataset\n",
-        "dataset = tfds.load(name=\"mnist\", split=tfds.Split.TRAIN)\n",
+        "dataset = tfds.load(name=\"mnist\", split=\"train\")\n",
         "\n",
         "# Build your input pipeline\n",
         "dataset = dataset.shuffle(1024).batch(32).prefetch(tf.data.experimental.AUTOTUNE)\n",

docs/_index.yaml

@@ -33,7 +33,7 @@ landing_page:
         print(tfds.list_builders())
 
         # Construct a tf.data.Dataset
-        dataset = tfds.load(name="mnist", split=tfds.Split.TRAIN)
+        dataset = tfds.load(name="mnist", split="train")
 
         # Build your input pipeline
         dataset = dataset.shuffle(1024).batch(32).prefetch(tf.data.experimental.AUTOTUNE)

docs/add_dataset.md

@@ -6,10 +6,10 @@ See our [list of datasets](catalog/overview.md) to see if the dataset you want
 isn't already added.
 
 *   [Overview](#overview)
-*   [Writing `my_dataset.py`](#writing-my-datasetpy)
+*   [Writing `my_dataset.py`](#writing-my_datasetpy)
     *   [Use the default template](#use-the-default-template)
-    *   [DatasetBuilder](#datasetbuilde)
-    *   [my_dataset.py](#my-datasetpy)
+    *   [DatasetBuilder](#datasetbuilder)
+    *   [my_dataset.py](#my_datasetpy)
 *   [Specifying `DatasetInfo`](#specifying-datasetinfo)
     *   [`FeatureConnector`s](#featureconnectors)
 *   [Downloading and extracting source data](#downloading-and-extracting-source-data)
@@ -558,6 +558,10 @@ for more information.
 
 Send the pull request for review.
 
+When creating the pull request, fill in the areas for the name, issue reference,
+and GitHub Gist link. When using the checklist, replace each `[ ]` with `[x]` to
+mark it off.
+
 
 ## Define the dataset outside TFDS.