The stacks of literal verse contributed has increased drastically in recent millennia, which creates a opulence of particulars for exploration and the extortion of knowledge. The consumption and processing of the ever-growing volume and complexity of text data produced daily, including Web broadcast, Editorial pieces, E-messages, and messaging are rendered challenging by people due to the overwhelming nature of the information. However, this also means that it can be prolonged and rigorous to manually sift over generous textual quantity to discern pertinent insights. Text summarization overcomes these hurdles by instinctively condensing text into a shorter, more manageable form, allowing users to rapidly acquire the Critical insights. Amid the prevailing data- centric conditions, this emerged as a vital tool that is applied in a variety of context like Analysis of insights, personalized content endorsements, and knowledge recuperation. This article proposes an Abstractive Text Summarization (ATS) framework which utilizes a combination of Seq2Seq model with Long Short-Term Memory (LSTM) and Convolutional Neural Network (CNN), bidirectional LSTM and Hybrid LSTM with attention to construct new sentences. This text summarization framework, referred as Deep Learning based Abstractive Text Summarization (ATSDL) which yields summarization from origin sentences using deep learning and is designed to explore more fine-grained fragments of text, such as semantic phrases, to create summaries. A massive corpus of text data called the News summary dataset is frequently used to train algorithms for multi-sentence summarizing tasks. To assess the quality of text produced by the deep learning model, two assessment metrics that are often employed in the natural language processing discipline are ROUGE and BLEU. Experimental results on news summary and CNN/DailyMail dataset revealed that ATSDL method achieves recent contemporary models in aspects of Grammatical structures and contextual relations, as well as having challenging outcomes in manual textual performance evaluations.