v6.0.0b4

This will probably be the final release, except for bug fixes.
Thanks to all who have expressed interest in vocal-remover!!!

Note

• Add evaluation script
• Refactor

v6.0.0b3

Note

• Phase estimation with complex mask
• Update the model architecture
• Update the pre-trained model

Evaluation result on DSD100

SDR in dB

Version	Instruments	Vocals
v4.0.0	15.778	5.214
v5.1.1	15.858	5.294
v6.0.0b3	16.030	5.614

v5.1.1

Note

• baseline.pth now can be loaded from anywhere (#138)
• Update requirements.txt

v6.0.0b2

Warning

This is a beta release.

Note

• Implement bounded mask
• Update the pre-trained model

v5.1.0

Note

• Fix a bug that causes a large disk load on online data loading

v6.0.0b1

Warning

This is a beta release.

Note

• Phase estimation with complex mask
• Update model architecture
• Update the pre-trained model
• Fix a bug that causes a large disk load on online data loading

Evaluation result on DSD100

Instruments

	SDR	ISR	SIR	SAR
v4.0.0	15.778	25.336	18.903	15.182
v5.0.5	15.858	26.580	18.740	15.640
v6.0.0b1	16.067	27.027	19.938	15.985

Vocals

	SDR	ISR	SIR	SAR
v4.0.0	5.214	9.985	11.796	5.426
v5.0.5	5.294	9.810	12.994	5.549
v6.0.0b1	5.503	10.903	13.817	6.018

v5.0.5

Note

• Add resampy to requirements.txt (#146)

v5.0.4

Note

• Update requirements.txt (#136)

v5.0.3

Note

• Update link to the paper (#113)
• Add support Apple chip device (#121)
• Updated librosa function calls that do not comply to PEP3102 (#125)

v5.0.2

Note

• Parameterize model size
• Add --output_dir option (#100)