Merge branch 'keras-team:master' into master

keras-team · Jun 18, 2023 · f396cc9 · f396cc9
2 parents 4fee4db + d0569ec
commit f396cc9
Show file tree

Hide file tree

Showing 150 changed files with 9,844 additions and 4,369 deletions.
diff --git a/.github/workflows/stale-issues-pr.yml b/.github/workflows/stale-issues-pr.yml
@@ -0,0 +1,46 @@
+name: Close inactive issues
+on:
+  schedule:
+    - cron: "30 1 * * *"
+
+jobs:
+  close-issues:
+    runs-on: ubuntu-latest
+    permissions:
+      issues: write
+      pull-requests: write
+    steps:
+      - name: Awaiting response issues
+        uses: actions/stale@v5
+        with:
+          days-before-issue-stale: 14
+          days-before-issue-close: 14
+          stale-issue-label: "stale"
+          close-issue-reason: completed
+          only-labels: "stat:awaiting response from contributor"
+          stale-issue-message: > 
+            This issue is stale because it has been open for 14 days with no activity.
+            It will be closed if no further activity occurs. Thank you.
+          close-issue-message: >
+            This issue was closed because it has been inactive for 28 days.
+            Please reopen if you'd like to work on this further.
+          days-before-pr-stale: 14
+          days-before-pr-close: 14
+          stale-pr-message: "This PR is stale because it has been open for 14 days with no activity. It will be closed if no further activity occurs. Thank you."
+          close-pr-message: "This PR was closed because it has been inactive for 28 days. Please reopen if you'd like to work on this further."
+          repo-token: ${{ secrets.GITHUB_TOKEN }}
+      - name: Contribution issues
+        uses: actions/stale@v5
+        with:
+          days-before-issue-stale: 180
+          days-before-issue-close: 365
+          stale-issue-label: "stale"
+          # reason for closed the issue default value is not_planned
+          close-issue-reason: not_planned
+          any-of-labels: "stat:contributions welcome,good first issue"
+          stale-issue-message: > 
+            This issue is stale because it has been open for 180 days with no activity.
+            It will be closed if no further activity occurs. Thank you.
+          close-issue-message: >
+            This issue was closed because it has been inactive for more than 1 year.
+          repo-token: ${{ secrets.GITHUB_TOKEN }}
diff --git a/README.md b/README.md
@@ -78,6 +78,7 @@ Author: (could be `Authors`: as well, and may contain markdown links)
 Date created: (date in yyyy/mm/dd format)
 Last modified: (date in yyyy/mm/dd format)
 Description: (one-line text description)
+Accelerator: (could be GPU, TPU, or None)
 ```
 
 To see examples of tutobooks, you can check out any `.py` file in `examples/` or `guides/`.

diff --git a/examples/audio/ctc_asr.py b/examples/audio/ctc_asr.py
@@ -187,7 +187,7 @@ def encode_single_sample(wav_file, label):
 """
 
 batch_size = 32
-# Define the trainig dataset
+# Define the training dataset
 train_dataset = tf.data.Dataset.from_tensor_slices(
     (list(df_train["file_name"]), list(df_train["normalized_transcription"]))
 )

diff --git a/examples/audio/ipynb/ctc_asr.ipynb b/examples/audio/ipynb/ctc_asr.ipynb
@@ -281,7 +281,7 @@
    "outputs": [],
    "source": [
     "batch_size = 32\n",
-    "# Define the trainig dataset\n",
+    "# Define the training dataset\n",
     "train_dataset = tf.data.Dataset.from_tensor_slices(\n",
     "    (list(df_train[\"file_name\"]), list(df_train[\"normalized_transcription\"]))\n",
     ")\n",
@@ -675,4 +675,4 @@
  },
  "nbformat": 4,
  "nbformat_minor": 0
-}
+}
diff --git a/examples/audio/ipynb/melgan_spectrogram_inversion.ipynb b/examples/audio/ipynb/melgan_spectrogram_inversion.ipynb
@@ -122,7 +122,7 @@
    "source": [
     "We create a `tf.data.Dataset` to load and process the audio files on the fly.\n",
     "The `preprocess()` function takes the file path as input and returns two instances of the\n",
-    "wave, one for input and one as the ground truth for comparsion. The input wave will be\n",
+    "wave, one for input and one as the ground truth for comparison. The input wave will be\n",
     "mapped to a spectrogram using the custom `MelSpec` layer as shown later in this example."
    ]
   },
@@ -287,7 +287,7 @@
     "    \"\"\"Convolutional residual stack with weight normalization.\n",
     "\n",
     "    Args:\n",
-    "        filter: int, determines filter size for the residual stack.\n",
+    "        filters: int, determines filter size for the residual stack.\n",
     "\n",
     "    Returns:\n",
     "        Residual stack output.\n",
@@ -860,4 +860,4 @@
  },
  "nbformat": 4,
  "nbformat_minor": 0
-}
+}
diff --git a/examples/audio/ipynb/speaker_recognition_using_cnn.ipynb b/examples/audio/ipynb/speaker_recognition_using_cnn.ipynb
@@ -92,8 +92,8 @@
     "SHUFFLE_SEED = 43\n",
     "\n",
     "# The sampling rate to use.\n",
-    "# This is the one used in all of the audio samples.\n",
-    "# We will resample all of the noise to this sampling rate.\n",
+    "# This is the one used in all the audio samples.\n",
+    "# We will resample all the noise to this sampling rate.\n",
     "# This will also be the output size of the audio wave samples\n",
     "# (since all samples are of 1 second long)\n",
     "SAMPLING_RATE = 16000\n",
@@ -214,7 +214,7 @@
     "In this section:\n",
     "\n",
     "- We load all noise samples (which should have been resampled to 16000)\n",
-    "- We split those noise samples to chuncks of 16000 samples which\n",
+    "- We split those noise samples to chunks of 16000 samples which\n",
     "correspond to 1 second duration each"
    ]
   },
@@ -647,4 +647,4 @@
  },
  "nbformat": 4,
  "nbformat_minor": 0
-}
+}
diff --git a/examples/audio/ipynb/transformer_asr.ipynb b/examples/audio/ipynb/transformer_asr.ipynb
@@ -72,7 +72,7 @@
     "position embeddings and token embeddings.\n",
     "\n",
     "When processing audio features, we apply convolutional layers to downsample\n",
-    "them (via convolution stides) and process local relationships."
+    "them (via convolution strides) and process local relationships."
    ]
   },
   {
@@ -610,7 +610,7 @@
     "            self.lr_after_warmup\n",
     "            - (epoch - self.warmup_epochs)\n",
     "            * (self.lr_after_warmup - self.final_lr)\n",
-    "            / (self.decay_epochs),\n",
+    "            / self.decay_epochs,\n",
     "        )\n",
     "        return tf.math.minimum(warmup_lr, decay_lr)\n",
     "\n",
@@ -720,4 +720,4 @@
  },
  "nbformat": 4,
  "nbformat_minor": 0
-}
+}
diff --git a/examples/audio/ipynb/uk_ireland_accent_recognition.ipynb b/examples/audio/ipynb/uk_ireland_accent_recognition.ipynb
@@ -102,7 +102,7 @@
     "# We see that there are 2 compressed files for each accent (except Irish):\n",
     "# - One for male speakers\n",
     "# - One for female speakers\n",
-    "# However, we will using a gender agnostic dataset.\n",
+    "# However, we will be using a gender agnostic dataset.\n",
     "\n",
     "# List of gender agnostic categories\n",
     "gender_agnostic_categories = [\n",
@@ -145,7 +145,6 @@
     "import os\n",
     "import io\n",
     "import csv\n",
-    "import random\n",
     "import numpy as np\n",
     "import pandas as pd\n",
     "import tensorflow as tf\n",
@@ -211,7 +210,7 @@
     "[Crowdsourced high-quality UK and Ireland English Dialect speech data set](https://openslr.org/83/)\n",
     "which consists of a total of 17,877 high-quality audio wav files.\n",
     "\n",
-    "This dataset includes over 31 hours of recording from 120 vounteers who self-identify as\n",
+    "This dataset includes over 31 hours of recording from 120 volunteers who self-identify as\n",
     "native speakers of Southern England, Midlands, Northern England, Wales, Scotland and Ireland.\n",
     "\n",
     "For more info, please refer to the above link or to the following paper:\n",
@@ -243,7 +242,7 @@
     "    fname=\"line_index_file\", origin=URL_PATH + \"line_index_all.csv\"\n",
     ")\n",
     "\n",
-    "# Download the list of compressed files that contains the audio wav files\n",
+    "# Download the list of compressed files that contain the audio wav files\n",
     "for i in zip_files:\n",
     "    fname = zip_files[i].split(\".\")[0]\n",
     "    url = URL_PATH + zip_files[i]\n",
@@ -1060,4 +1059,4 @@
  },
  "nbformat": 4,
  "nbformat_minor": 0
-}
+}
diff --git a/examples/audio/ipynb/wav2vec2_audiocls.ipynb b/examples/audio/ipynb/wav2vec2_audiocls.ipynb
@@ -296,16 +296,16 @@
    "source": [
     "Before we can feed the audio utterance samples to our model, we need to\n",
     "pre-process them. This is done by a Hugging Face Transformers \"Feature Extractor\"\n",
-    "which will (as the name indicates) re-sample your the inputs to sampling rate\n",
-    "the the model expects (in-case they exist with a different sampling rate), as well\n",
+    "which will (as the name indicates) re-sample your inputs to the sampling rate\n",
+    "the model expects (in-case they exist with a different sampling rate), as well\n",
     "as generate the other inputs that model requires.\n",
     "\n",
     "To do all of this, we instantiate our `Feature Extractor` with the\n",
     "`AutoFeatureExtractor.from_pretrained`, which will ensure:\n",
     "\n",
     "We get a `Feature Extractor` that corresponds to the model architecture we want to use.\n",
     "We download the config that was used when pretraining this specific checkpoint.\n",
-    "This will be cached so it's not downloaded again the next time we run the cell.\n",
+    "This will be cached so that it's not downloaded again the next time we run the cell.\n",
     "\n",
     "The `from_pretrained()` method expects the name of a model from the Hugging Face Hub. This is\n",
     "exactly similar to `MODEL_CHECKPOINT` and we just pass that.\n",
@@ -373,7 +373,7 @@
    },
    "source": [
     "We now define our model. To be precise, we define a Wav2Vec 2.0 model and add a\n",
-    "Classification-Head on top to output a probability ditribution of all classes for each\n",
+    "Classification-Head on top to output a probability distribution of all classes for each\n",
     "input audio sample. Since the model might get complex we first define the Wav2Vec\n",
     "2.0 model with Classification-Head as a Keras layer and then build the model using that.\n",
     "\n",
@@ -574,7 +574,7 @@
     "Great! Now that we have trained our model, we predict the classes\n",
     "for audio samples in the test set using the `model.predict()` method! We see\n",
     "the model predictions are not that great as it has been trained on a very small\n",
-    "number of samples for just 1 epoch. For best results, we reccomend training on\n",
+    "number of samples for just 1 epoch. For best results, we recommend training on\n",
     "the complete dataset for at least 5 epochs!"
    ]
   },
@@ -623,7 +623,7 @@
     "colab_type": "text"
    },
    "source": [
-    "Now you can push this model to Hugging Face Model Hub and also share it with with all your friends,\n",
+    "Now you can push this model to Hugging Face Model Hub and also share it with all your friends,\n",
     "family, favorite pets: they can all load it with the identifier\n",
     "`\"your-username/the-name-you-picked\"`, for instance:\n",
     "\n",

diff --git a/examples/audio/md/ctc_asr.md b/examples/audio/md/ctc_asr.md
@@ -261,7 +261,7 @@ appeared in the input.
 
 ```python
 batch_size = 32
-# Define the trainig dataset
+# Define the training dataset
 train_dataset = tf.data.Dataset.from_tensor_slices(
     (list(df_train["file_name"]), list(df_train["normalized_transcription"]))
 )

diff --git a/examples/audio/md/melgan_spectrogram_inversion.md b/examples/audio/md/melgan_spectrogram_inversion.md
@@ -107,7 +107,7 @@ LJSpeech-1.1.tar.bz 100%[===================>]   2.56G  68.3MB/s    in 36s
 
 We create a `tf.data.Dataset` to load and process the audio files on the fly.
 The `preprocess()` function takes the file path as input and returns two instances of the
-wave, one for input and one as the ground truth for comparsion. The input wave will be
+wave, one for input and one as the ground truth for comparison. The input wave will be
 mapped to a spectrogram using the custom `MelSpec` layer as shown later in this example.
 
 
@@ -240,7 +240,7 @@ def residual_stack(input, filters):
     """Convolutional residual stack with weight normalization.
 
     Args:
-        filter: int, determines filter size for the residual stack.
+        filters: int, determines filter size for the residual stack.
 
     Returns:
         Residual stack output.

diff --git a/examples/audio/md/speaker_recognition_using_cnn.md b/examples/audio/md/speaker_recognition_using_cnn.md
@@ -70,8 +70,8 @@ VALID_SPLIT = 0.1
 SHUFFLE_SEED = 43
 
 # The sampling rate to use.
-# This is the one used in all of the audio samples.
-# We will resample all of the noise to this sampling rate.
+# This is the one used in all the audio samples.
+# We will resample all the noise to this sampling rate.
 # This will also be the output size of the audio wave samples
 # (since all samples are of 1 second long)
 SAMPLING_RATE = 16000
@@ -167,7 +167,7 @@ for folder in os.listdir(DATASET_ROOT):
 In this section:
 
 - We load all noise samples (which should have been resampled to 16000)
-- We split those noise samples to chuncks of 16000 samples which
+- We split those noise samples to chunks of 16000 samples which
 correspond to 1 second duration each
 
 

diff --git a/examples/audio/md/transformer_asr.md b/examples/audio/md/transformer_asr.md
@@ -51,7 +51,7 @@ When processing past target tokens for the decoder, we compute the sum of
 position embeddings and token embeddings.
 
 When processing audio features, we apply convolutional layers to downsample
-them (via convolution stides) and process local relationships.
+them (via convolution strides) and process local relationships.
 
 
 ```python
@@ -511,7 +511,7 @@ class CustomSchedule(keras.optimizers.schedules.LearningRateSchedule):
             self.lr_after_warmup
             - (epoch - self.warmup_epochs)
             * (self.lr_after_warmup - self.final_lr)
-            / (self.decay_epochs),
+            / self.decay_epochs,
         )
         return tf.math.minimum(warmup_lr, decay_lr)