Search results

MERT configuration in fairseq seems to differ from the configuration in huggingface

Status: Open.

#20 In yizhilll/MERT;

· LiableFishYS opened on Jan 17, 2025

Why MERT-v1-330M was trained with feat_extract_norm="group" instead of "layer"?

Status: Open.

#19 In yizhilll/MERT;

· LiableFishYS opened on Jan 17, 2025

Fine-tune MERT model on the custom dataset

Status: Open.

#18 In yizhilll/MERT;

· kv-42 opened on Dec 26, 2024

HuBERT and data2vec retrained on music

Status: Open.

#17 In yizhilll/MERT;

· lcc-404 opened on Aug 25, 2024

Music Descriptor prediction (especially for EMO task)

Status: Open.

#16 In yizhilll/MERT;

· uu95 opened on Jun 20, 2024

Questions about data preparation

Status: Open.

#15 In yizhilll/MERT;

· nkkanee opened on Apr 23, 2024

Reproducing the numbers

Status: Open.

#13 In yizhilll/MERT;

· Laubeee opened on Dec 26, 2023

Regarding attention relaxation

Status: Open.

#10 In yizhilll/MERT;

· PetrosVav opened on Nov 15, 2023

questions about dictionary.encode_line to process encodec labels

Status: Open.

#8 In yizhilll/MERT;

· xujinchang opened on Oct 16, 2023

High vram usage for long audio.

Status: Open.

#6 In yizhilll/MERT;

· hykilpikonna opened on Aug 11, 2023