Issues

• Mert

  yizhilll/MERT#22 · Filter by author Cloud-mert opened on Sep 18, 2025

• Why use Conv1d(k=1) as the layer aggregator in the inference demo?

  yizhilll/MERT#21 · Filter by author kmraven opened on Aug 26, 2025

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

  yizhilll/MERT#20 · Filter by author LiableFishYS opened on Jan 17, 2025

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

  yizhilll/MERT#19 · Filter by author LiableFishYS opened on Jan 17, 2025
  1

• Fine-tune MERT model on the custom dataset

  yizhilll/MERT#18 · Filter by author kv-42 opened on Dec 26, 2024

• HuBERT and data2vec retrained on music

  yizhilll/MERT#17 · Filter by author lcc-404 opened on Aug 25, 2024
  1

• Music Descriptor prediction (especially for EMO task)

  yizhilll/MERT#16 · Filter by author uu95 opened on Jun 20, 2024

• Questions about data preparation

  yizhilll/MERT#15 · Filter by author nkkanee opened on Apr 23, 2024

• Reproducing the numbers

  yizhilll/MERT#13 · Filter by author Laubeee opened on Dec 26, 2023
  1

• Regarding attention relaxation

  yizhilll/MERT#10 · Filter by author PetrosVav opened on Nov 15, 2023

• questions about dictionary.encode_line to process encodec labels

  yizhilll/MERT#8 · Filter by author xujinchang opened on Oct 16, 2023
  7

• High vram usage for long audio.

  yizhilll/MERT#6 · Filter by author hykilpikonna opened on Aug 11, 2023
  1