Update datasets task tags to align tags with models (#4067)

* update tasks list

* update tags in dataset cards

* more cards updates

* update dataset tags parser

* fix multi-choice-qa

* style

* small improvements in some dataset cards

* allow certain tag fields to be empty

* update vision datasets tags

* use multi-class-image-classification and remove other tags

Commit from https://github.com/huggingface/datasets/commit/edb4411d4e884690b8b328dba4360dbda6b3cbc8

README.md

@@ -17,9 +17,10 @@ size_categories:
 source_datasets:
 - original
 task_categories:
-- speech-processing
-task_ids:
 - automatic-speech-recognition
+- audio-classification
+task_ids:
+- audio-speaker-identification
 ---
 
 # Dataset Card for librispeech_asr
@@ -62,7 +63,7 @@ LibriSpeech is a corpus of approximately 1000 hours of 16kHz read English speech
 
 ### Supported Tasks and Leaderboards
 
-- `automatic-speech-recognition`, `speaker-identification`: The dataset can be used to train a model for Automatic Speech Recognition (ASR). The model is presented with an audio file and asked to transcribe the audio file to written text. The most common evaluation metric is the word error rate (WER). The task has an active leaderboard which can be found at https://paperswithcode.com/sota/speech-recognition-on-librispeech-test-clean and ranks models based on their WER.
+- `automatic-speech-recognition`, `audio-speaker-identification`: The dataset can be used to train a model for Automatic Speech Recognition (ASR). The model is presented with an audio file and asked to transcribe the audio file to written text. The most common evaluation metric is the word error rate (WER). The task has an active leaderboard which can be found at https://paperswithcode.com/sota/speech-recognition-on-librispeech-test-clean and ranks models based on their WER.
 
 ### Languages