SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2407.09099 · 2024

Music Proofreading with RefinPaint: Where and How to Modify Compositions given Context

arXiv · PDF · Open in the Atlas

Code that ran

We lifted 11 functions out of this paper's own repositories and ran 10 of them in a sandbox. "Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results.

RepositoryRoleRan
ta603/refinpaint canonical 10 of 11
FunctionStatusWhere it lives
Linear Ran ta603/refinpaint/EfficcientTransformer.py
pointer only (licence: NONE) · get_code("01f0f6d4a6e475b1")
data_augmentation_tokens Ran ta603/refinpaint/MidiTok_modified/miditok/data_augmentation/data_augmentation.py
pointer only (licence: NOASSERTION) · get_code("56a5d46981788af1")
is_more_or_less_red Ran ta603/refinpaint/visual.py
pointer only (licence: NOASSERTION) · get_code("2ed59da4722ce8f2")
load_binary_data Ran ta603/refinpaint/RefinePaint.py
pointer only (licence: NOASSERTION) · get_code("8611c5ea60acf2e7")
load_json Ran ta603/refinpaint/FeedbackModel.py
pointer only (licence: NONE) · get_code("593e5e4d03d4d4df")
load_model_state_from_parts Ran ta603/refinpaint/RefinePaint.py
pointer only (licence: NOASSERTION) · get_code("1fbead6df1170d53")
modify_labels_avg_pool Ran ta603/refinpaint/FeedbackModel.py
pointer only (licence: NOASSERTION) · get_code("d1469ea687fe7680")
scaled_attention Ran ta603/refinpaint/EfficcientTransformer.py
pointer only (licence: NOASSERTION) · get_code("511c979f2952f219")
top_k Ran ta603/refinpaint/InpaintingModel.py
pointer only (licence: NONE) · get_code("39f3cec3c6be459f")
top_p Ran ta603/refinpaint/InpaintingModel.py
pointer only (licence: NOASSERTION) · get_code("134ecdfe2b0dc9f1")
top_p_sampling Not yet run ta603/refinpaint/EfficcientTransformer.py
pointer only (licence: NOASSERTION) · get_code("77d5cc275e2c2529")

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

Autoregressive generative transformers are key in music generation, producing coherent compositions but facing challenges in human-machine collaboration. We propose RefinPaint, an iterative technique that improves the sampling process. It does this by identifying the weaker music elements using a feedback model, which then informs the choices for resampling by an inpainting model. This dual-focus methodology not only facilitates the machine's ability to improve its automatic inpainting generation through repeated cycles but also offers a valuable tool for humans seeking to refine their compositions with automatic proofreading. Experimental results suggest RefinPaint's effectiveness in inpainting and proofreading tasks, demonstrating its value for refining music created by both machines and humans. This approach not only facilitates creativity but also aids amateur composers in improving their work.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("2407.09099")
get_code_for_paper("2407.09099")
have("2407.09099")

Connect an agent — have() is free.