Thomas Pellegrini, Ludovic Tuncay, Étienne Labb É Id
We lifted 1 functions out of this paper's own repositories and ran 1 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.
| Repository | Role | Ran |
|---|---|---|
| LudovicTuncay/audio-embeddings | canonical | 1 of 1 |
| Function | Status | Where it lives |
|---|---|---|
| convert_checkpoint_to_safetensors | Ran | LudovicTuncay/audio-embeddings/src/callbacks/safetensors_callback.py code served (permissive licence) · get_code("4d4e38c9c5491268") |
Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.
Self-supervised learning enables audio representations that transfer across domains and tasks. We present BEST-RQ-2, an evolution of BEST-RQ that retains frozen randomprojection-based discrete targets while introducing a two-step contextualize-then-predict pretraining scheme. A ViT context encoder processes only the unmasked spectrogram regions, and a lightweight predictor infers targets for the masked regions; the predictor is discarded after pretraining. Replacing the original Conformer encoder with a ViT shifts performance across domains, slightly reducing speech performance while improving music and environmental sounds, with comparable average scores. The main improvement comes from decomposing masked prediction into separate contextualization and prediction stages. On the X-ARES and XARES-LLM benchmarks, BEST-RQ-2 consistently outperforms one-stage baselines in overall transfer while keeping inference compute unchanged. Code 1 and model checkpoints 2 are publicly available.
The same record, over MCP at https://syntology.ai/mcp:
get_harvested_code_for_paper("2606.30700")
get_code_for_paper("2606.30700")
have("2606.30700")
Connect an agent — have() is free.