{"type":"video","version":"1.0","html":"<iframe src=\"https://www.loom.com/embed/cf7cee8151544d6c86467db16ecb858f\" frameborder=\"0\" width=\"1280\" height=\"960\" webkitallowfullscreen mozallowfullscreen allowfullscreen></iframe>","height":960,"width":1280,"provider_name":"Loom","provider_url":"https://www.loom.com","thumbnail_height":960,"thumbnail_width":1280,"thumbnail_url":"https://cdn.loom.com/sessions/thumbnails/cf7cee8151544d6c86467db16ecb858f-00001.gif","duration":88.557,"title":"Developing a Custom DistilBERT Model for Semantic Classification","description":"In this video, I discuss my project of developing a custom DistilBERT model with a 768-dimensional hidden layer for semantic classification tasks. I explain what semantic classification is and how it can be used for sentiment analysis, topic categorization, intent detection, and spam detection. I also share my goal of optimizing the performance of the DistilBERT model for specific needs and the tools I used, such as the Intel Developer Cloud Notebook and the Transformers library from Hugging Face. No action is requested from the viewers."}