hybrid attention deep leaning model for text classification

I am looking for a skilled freelancer who can build a hybrid attention deep learning model for text classification. The project involves the following requirements:

Data: The model will be classifying social media posts.

Programming Language: The preferred language for this project is Python.

Accuracy: The desired accuracy level for the text classification model is 90-95%.

Ideal Skills and Experience:

- Experience in building deep learning models for text classification
- Knowledge of natural language processing (NLP) techniques and tools
- Proficiency in Python programming language
- Familiarity with attention mechanisms and hybrid models
- Ability to work with large datasets
1- Prepare a dataset of four dialects Egyptian, Gulf, Yemen, and Jordan. You can extract them from social media, YouTube.
2- Pre-processing steps in order to improve the model accuracy
3- Draw the model Architecture
4- Use TF-IDF , Word2Vec, Glove , and compare the accuracy, F1-Svore, precision , recall.
5- Use two BiLSTM models, or LSTM with the Attention model.
6- Write the result and discussion
If you possess the necessary skills and experience, please submit your proposal. Thank you.