1

app.py

@@ -117,7 +117,7 @@ def engine(text_input):
 app = gr.Interface(engine, 
                    gr.inputs.Textbox(lines=5, label="Input Text"),
                    gr.outputs.Video(type=None, label='Final Merged video'), 
-                   description="<div>Firstly, the Demo generates speech from input-text using facebook's fastspeech2-en-ljspeech from <a href='https://huggingface.co/facebook/fastspeech2-en-ljspeech' target='_blank'>HF hub</a>.<br>Then, takes the input-text and extracts the entities in it using Flair NER model from <a href='https://huggingface.co/flair/ner-english-ontonotes-large' target='_blank'>HF Hub</a>. <br>Then, generate images using <a href='https://huggingface.co/spaces/multimodalart/latentdiffusion' target='_blank'>Multimodalart Space</a> for every entity separately.<br>Creates a video by stringing all the entity-images together. <br>Lastly, Fuses the AI enerated audio and video together to create a coherent movie for you to watch. <br><br>A fun little app that lets you turn your text to video (well, in some ways atleast :) ). More the entities in your text, More time to build the output (expect 10-20 seconds per entity), More fun to watch.</div>" ,
+                   description="<div>Firstly, the Demo generates speech from input-text using facebook's fastspeech2-en-ljspeech from <a href='https://huggingface.co/facebook/fastspeech2-en-ljspeech' target='_blank'>HF hub</a>.<br>Then, takes the input-text and extracts the entities in it using Flair NER model from <a href='https://huggingface.co/flair/ner-english-ontonotes-large' target='_blank'>HF Hub</a>. <br>Then, generate images using <a href='https://huggingface.co/spaces/multimodalart/latentdiffusion' target='_blank'>Multimodalart Space</a> for every entity separately.<br>Creates a video by stringing all the entity-images together. <br>Lastly, Fuses the AI generated audio and video together to create a coherent movie for you to watch. <br><br>A fun little app that lets you turn your text to video (well, in some ways atleast :) ). More the entities in your text, More time to build the output, More fun to watch.<br> Please expect build time of around 10-20 seconds per entity. For instance, in the first example there are 13 entities as per the NER model used here.</div>" ,
                    examples=["On April 17th Sunday George celebrated Easter. He is staying at Empire State building with his parents. He is a citizen of Canada and speaks English and French fluently. His role model is former president Obama. He got 1000 dollar from his mother to visit Disney World and to buy new iPhone mobile.  George likes watching Game of Thrones.", "April is the month of Easter weekend. Visit places like Statue of Liberty with friends. Take at least 200 dollars in cash with you. Use Android phone to find places in Newyork City."],
                    title="Generate Video from Text",
                    article="<br><div>For best results, make sure to enter a text that has entities listed on model card for <a href='https://huggingface.co/flair/ner-english-ontonotes-large' target='_blank'>flair/ner-english-ontonotes-large</a>. Some examples of type of entities that will be helpful are - Date values, event names, building names, languages, locations, money value, organization names, famous people names, products and so on.</div><br><h4 style='font-size: 110%;margin-top:1em'>Who owns the videos produced by this demo?</h4><div><i>(Borrowing this from multimodalart spaces)</i> Definetly not me! Probably you do. I say probably because the Copyright discussion about AI generated art is ongoing. So <a href='https://www.theverge.com/2022/2/21/22944335/us-copyright-office-reject-ai-generated-art-recent-entrance-to-paradise' target='_blank'>it may be the case that everything produced here falls automatically into the public domain</a>. But in any case it is either yours or is in the public domain.</div>"