24 days ago
-
Clarify pretraining and fine-tuning with standard FM
-
Remove obsolete private-access note from quick start
-
Link one shared source audio and preparation guide for training workflows
-
Display the complete paper title alongside the SteerSound model name
-
Align data workflow names and explicitly include user data
-
Clarify standard FM Complex evaluation and STEER data preparation workflows
-
Link both downstream fine-tuning workflows and update card checksum
27 days ago
-
Express model parameter counts in billions
-
Simplify model size and total inference memory requirements
-
Present final SteerSound weights and credit pretrained components
-
Document train-only targets in the published dataset
-
Add files using upload-large-folder tool