Added several typo fixes

Language-Scaling/German/README.md

@@ -22,7 +22,7 @@ When adapting Riva to a whole new language, a large amount of high-quality trans
 
 For German, there are several significant sources of public datasets that we can readily leverage:
 
-- [Mozila Common Voice](https://commonvoice.mozilla.org/en/datasets) (MCV) corpus 7.0, `DE` subset: 571 hours, ~ 26 Gbs.
+- [Mozilla Common Voice](https://commonvoice.mozilla.org/en/datasets) (MCV) corpus 7.0, `DE` subset: 571 hours, ~ 26 Gbs.
 - [Multilingual LibriSpeech](http://www.openslr.org/94/) (MLS), `DE` subset: 1918 hours, ~115 GBs.
 - [Voxpopuli](https://ai.facebook.com/blog/voxpopuli-the-largest-open-multilingual-speech-corpus-for-ai-translation-and-more/), `DE` subset: 214 hours, 4.6 Gbs.
 

Language-Scaling/German/data_preparation/README.md

@@ -9,7 +9,7 @@ When adapting Riva to a whole new language, a large amount of high-quality trans
 
 For German, there are several significant sources of public datasets that we can readily leverage:
 
-- [Mozila Common Voice](https://commonvoice.mozilla.org/en/datasets) (MCV) corpus 7.0, `DE` subset: 571 hours 
+- [Mozilla Common Voice](https://commonvoice.mozilla.org/en/datasets) (MCV) corpus 7.0, `DE` subset: 571 hours 
 - [Multilingual LibriSpeech](http://www.openslr.org/94/) (MLS), `DE` subset: 1918 hours
 - [VoxPopuli](https://ai.facebook.com/blog/voxpopuli-the-largest-open-multilingual-speech-corpus-for-ai-translation-and-more/), `DE` subset: 214 hours
 

Language-Scaling/German/data_preparation/scripts/process_asr_text_tokenizer.py

@@ -12,7 +12,7 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-# USAGE: python process_asr_text_tokenizer.py --manifest=<path to train manifest files, seperated by commas> \
+# USAGE: python process_asr_text_tokenizer.py --manifest=<path to train manifest files, separated by commas> \
 #         --data_root="<output directory>" \
 #         --vocab_size=<number of tokens in vocabulary> \
 #         --tokenizer=<"spe" or "wpe"> \
@@ -45,7 +45,7 @@
 #   --tokenizer: Can be either spe or wpe . spe refers to the Google sentencepiece library tokenizer.
 #       wpe refers to the HuggingFace BERT Word Piece tokenizer.
 #
-#   --no_lower_case: When this flag is passed, it will force the tokenizer to create seperate tokens for
+#   --no_lower_case: When this flag is passed, it will force the tokenizer to create separate tokens for
 #       upper and lower case characters. By default, the script will turn all the text to lower case
 #       before tokenization (and if upper case characters are passed during training/inference, the
 #       tokenizer will emit a token equivalent to Out-Of-Vocabulary). Used primarily for the
@@ -65,7 +65,7 @@
 #       positive integer. By default, any negative value (default = -1) will use the entire dataset.
 #
 #   --spe_train_extremely_large_corpus: When training a sentencepiece tokenizer on very large amounts of text,
-#       sometimes the tokenizer will run out of memory or wont be able to process so much data on RAM.
+#       sometimes the tokenizer will run out of memory or won't be able to process so much data on RAM.
 #       At some point you might receive the following error - "Input corpus too large, try with
 #       train_extremely_large_corpus=true". If your machine has large amounts of RAM, it might still be possible
 #       to build the tokenizer using the above flag. Will silently fail if it runs out of RAM.

Language-Scaling/German/training/1-train-acoustic-model.ipynb

@@ -17,7 +17,7 @@
     "\n",
     "![png](./imgs/german-transfer-learning.PNG)\n",
     "\n",
-    "We first demonstrate the training process with NeMo on 1 GPU in this notebook. To speed up training, multiple GPUs should be leveraged using the more efficient DDP (distributed data parallel) protocol, which must run in a seperate [training script](./train.py).\n",
+    "We first demonstrate the training process with NeMo on 1 GPU in this notebook. To speed up training, multiple GPUs should be leveraged using the more efficient DDP (distributed data parallel) protocol, which must run in a separate [training script](./train.py).\n",
     "\n",
     "This notebook can be run from within the NeMo container, such as:\n",
     "\n",

Language-Scaling/Hindi/README.md

@@ -139,7 +139,7 @@ In addition to evaluating our model on the train test split, we've also evaluate
 * [Hindi subtask-1](https://us.openslr.org/resources/103/subtask1_blindtest_wReadme.tar.gz) from [Interspeech MUCS Challenge](https://navana-tech.github.io/MUCS2021/challenge_details.html)
 * etc.
 
-We've observed very competetive WER, as low as 12.78, on these blind test sets.
+We've observed very competitive WER, as low as 12.78, on these blind test sets.
 
 ## 5. Riva Deployment
 

asr-improve-recognition-for-specific-words.md

@@ -51,7 +51,7 @@ Word boosting provides a quick and temporary, on-the-spot adaptation for the mod
 
 You will have to explicitly specify the list of boosted words at every request. Other adaptation methods such as custom vocabulary and lexicon mapping provide a more permanent solution, which affects every subsequent request.
 
-Pay attention to the followings while implementing word boosting:
+Pay attention to the following while implementing word boosting:
 - Word boosting can improve the chance of recognition of the desired words, but at the same time can increase false positives. As such, start with a small positive weight and gradually increase till you see positive effects. As a general guide, start with a boosted score of 20 and increase up to 100 if needed.
 - Word boosting is most suitable as a temporary fix for a new situation. However, if you wish to use it as a permanent adaptation, you can attempt binary search for the boosted weights while monitoring the accuracy metrics on a test set. The accuracy metrics should include both the word error rate (WER) and/or a form of term error rate (TER) focusing on the terms of interest.
 - Word Boosting is supported only with flashlight decoder.

deploy-aks.md

@@ -2,7 +2,7 @@
 
 # How do I Deploy Riva at Scale on Azure Cloud with AKS?
 
-This is an example of deploying and scaling Riva Speech Skills on Azure Cloud's Azure Kuberenetes Service (AKS)
+This is an example of deploying and scaling Riva Speech Skills on Azure Cloud's Azure Kubernetes Service (AKS)
 with Traefik-based load balancing. It includes the following steps:
 1. Creating the AKS cluster
 2. Deploying the Riva API service

deploy-gke.md

@@ -2,7 +2,7 @@
 
 # How do I Deploy Riva at Scale on Google Cloud with GKE?
 
-This is an example of deploying and scaling Riva Speech Skills on Google Cloud (GCP) Google Kuberenetes Engine (GKE)
+This is an example of deploying and scaling Riva Speech Skills on Google Cloud (GCP) Google Kubernetes Engine (GKE)
 with Traefik-based load balancing. It includes the following steps:
 1. Creating the GKE cluster
 2. Deploying the Riva API service