The Meta Pocket bookLlama AI instrument makes use of three massive language fashions to generate audio podcasts from blocks of textual content. Currently, the instrument solely accepts PDF information as enter, so customers should convert no matter textual content format they’ve into PDF.
Meta Pocket bookLlama workflow
Photo Credit: Meta
Pocket bookLlama first makes use of Llama 3.2 1B instruct mannequin to pre-process the PDF file and reserve it in a ‘.txt’ file. Then the Llama 3.1 70B instruct mannequin is used to jot down a podcast transcript utilizing the supply dataset. The transcription is then dramatised utilizing a re-writer which makes use of the Llama 3.1 8B instruct mannequin. Finally, a customized instrument is used so as to add the transcription in a text-to-speech workflow. For this, Meta is utilizing the Parler TTS instrument. Interested people can entry all of the fashions required to generate podcasts from the GitHub itemizing here.
However, the AI fashions talked about above are simply suggestions from the builders. Users can choose to make use of smaller fashions for each step, nonetheless, the outcomes might fluctuate. Meta highlighted that to run the AI system within the really helpful setup, customers would require a GPU with an aggregated reminiscence of roughly 140GB.
An X (previously often called Twitter) consumer posted a pattern of the generated podcast. Based on this, it seems the audio high quality is inferior to the Google NotebookLM, and it sounds shrill and robotic. Further, there are situations the place elements of audio get disregarded and the AI hosts find yourself talking over one another.
Meta acknowledges among the points and plans to enhance them within the subsequent iteration of the AI product. The firm highlighted, “The TTS mannequin is the limitation of how pure it will sound. This most likely be improved with a greater pipeline and with the assistance of somebody extra educated.”
The tech large can also be planning to make use of two totally different LLMs to jot down the script, the place every mannequin will debate the opposite to make the podcast sound extra conversational. This can also be a part of the builders’ future pipeline. Additionally, the corporate can also be testing the Llama 405B AI mannequin to jot down the transcripts in addition to growing the assist for extra enter and output codecs.
For the newest tech information and critiques, comply with Gadgets 360 on X, Facebook, WhatsApp, Threads and Google News. For the newest movies on devices and tech, subscribe to our YouTube channel. If you wish to know the whole lot about high influencers, comply with our in-house Who’sThat360 on Instagram and YouTube.
iPhone SE 4 Tipped to Arrive With 6.06-inch LTPS OLED Screen, 3,279mAh Battery and Apple’s In-House Modem
iPhone 16 Series Sales in Indonesia Blocked Over Unmet Investment Requirements
