<b>Multimodal Multilingual Hate Speech11K (MMHS11K) Dataset</b>

The Multimodal Multilingual Hate Speech 11K (MMHS11K) Dataset consists of 11,000 labeled Urdu Tweets with English translation, each categorized as either 'Hate' and 'No-Hate'. The dataset is divided into training (MMHS11K_train.xlsx ) and test (MMHS11K_test.xlsx ) files. The training file contains 8,800 records with an equal distribution (i.e. 4400) of 'Hate' and 'No-Hate' classes. The test file contains 2,200 records with 1,100 'Hate' and 1,100 'No-Hate' sample. Each file contains six columns:Tweet_Id: Represents the unique identifier of the corresponding tweet.Text: Contains the textual content of the tweet in Urdu.Text (English): Provides the English translation of the corresponding Urdu tweet.Image_Text: Displays text extracted from images in Urdu. If no text is extracted, the value 'NIL' is used.Image_Text (English): Contains the English translation of the corresponding Image_Text value. If Image_Text is 'NIL', this column also shows 'NIL'.Label: Indicates the classification of the multimodal multilingual tweet as either 'Hate' or 'No-Hate'The dataset contains two folders:MMHS11K_RGB_train.rar: This folder holds the training images in RGB format. It contains two subfolders: Hate and No-Hate. For each tweet in the MMHS11K_train.xlsx file, the corresponding image is stored in the appropriate subfolder (Hate or No-Hate) within the MMHS11K_RGB_train folder, using the tweet's Tweet_Id as the filename. For example, the image for Tweet_Id = 1596192893743796224 is stored in the No-Hate subfolder of MMHS11K_RGB_train, as this tweet is labeled as No-Hate.MMHS11K_RGB_test.rar: This folder contains the test images in RGB format. It also includes two subfolders: Hate and No-Hate.