Get our free extension to see links to code for papers anywhere online!

Add to Chrome

Add to Firefox

Picture for John Thickstun

John Thickstun

Melody transcription via generative pre-training

Dec 04, 2022
Chris Donahue, John Thickstun, Percy Liang

Figure 1 for Melody transcription via generative pre-training

Figure 2 for Melody transcription via generative pre-training

Figure 3 for Melody transcription via generative pre-training

Figure 4 for Melody transcription via generative pre-training

* Published as a conference paper at ISMIR 2022

Access Paper or Ask Questions

Parallel and Flexible Sampling from Autoregressive Models via Langevin Dynamics

May 17, 2021
Vivek Jayaram, John Thickstun

Figure 1 for Parallel and Flexible Sampling from Autoregressive Models via Langevin Dynamics

Figure 2 for Parallel and Flexible Sampling from Autoregressive Models via Langevin Dynamics

Figure 3 for Parallel and Flexible Sampling from Autoregressive Models via Langevin Dynamics

Figure 4 for Parallel and Flexible Sampling from Autoregressive Models via Langevin Dynamics

* 16 pages, 8 figures, to appear in ICML 2021

Access Paper or Ask Questions

Source Separation with Deep Generative Priors

Feb 19, 2020
Vivek Jayaram, John Thickstun

Figure 1 for Source Separation with Deep Generative Priors

Figure 2 for Source Separation with Deep Generative Priors

Figure 3 for Source Separation with Deep Generative Priors

Figure 4 for Source Separation with Deep Generative Priors

* 18 pages, 15 figures

Access Paper or Ask Questions

Convolutional Composer Classification

Nov 26, 2019
Harsh Verma, John Thickstun

Figure 1 for Convolutional Composer Classification

Figure 2 for Convolutional Composer Classification

Figure 3 for Convolutional Composer Classification

Figure 4 for Convolutional Composer Classification

* 8 pages, published at ISMIR 2019

Access Paper or Ask Questions