| Website | https://ondrej.cifka.com |
| GitHub | https://cifkao.github.io |
| https://twitter.com/cifkao |
🚀 We’re looking for a Master’s student to join our research team @ audioshake.ai for a 6-month internship!
Deep dive into PyTorch, optimize our SOTA audio models, and help make ML sound better (and faster) 🎶
Based in Paris or remote 🇫🇷 → https://audioshake.notion.site/Internship-ML-Optimization-205b133ddefe8025a9f2de74d30d4d38 #AudioML #Internship
Writing down lyrics for human consumption involves not only accurately capturing word sequences, but also incorporating punctuation and formatting for clarity and to convey contextual information. This includes song structure, emotional emphasis, and contrast between lead and background vocals. While automatic lyrics transcription (ALT) systems have advanced beyond producing unstructured strings of words and are able to draw on wider context, ALT benchmarks have not kept pace and continue to focus exclusively on words. To address this gap, we introduce Jam-ALT, a comprehensive lyrics transcription benchmark. The benchmark features a complete revision of the JamendoLyrics dataset, in adherence to industry standards for lyrics transcription and formatting, along with evaluation metrics designed to capture and assess the lyric-specific nuances, laying the foundation for improving the readability of lyrics. We apply the benchmark to recent transcription systems and present additional error analysis, as well as an experimental comparison with a classical music dataset.
So excited to finally share this really interesting project I worked on last year! 🐾 We trained a Transformer on animal trajectories (world-wide GPS location data for >50 species 🦆🦅🦓🦬🐢) and studied how movement history and environmental variables affect predictions.
https://doi.org/10.1101/2023.03.05.531080
https://github.com/cifkao/moveformer
Specifically, to compute the output distributions for all positions in a text of length N and all context lengths up to a max length C, we just need to run inference along a sliding window of length C, i.e. do N forward passes on segments of length ≤C. (see the illustration in my previous post)
Notice that this is a lot like generating a new sequence from the model (the naïve way)! 🧵4/4
The technique works with any causal LM, as long as it was trained to accept arbitrary text fragments (not necessarily starting at sentence or document boundary), which happens to be how large #GPT-like models (#GPT2, #GPT3, #GPTJ, ...) are usually trained.
The main trick is in realizing that the necessary probabilities can be computed efficiently by running the model along a sliding window. 🧵3/4
faro
