hyperparams.txt

	audio hidden	video hidden	text hidden	audio dropout	video dropout	text dropout	factor lr	learning rate	rank	batch size	weight decay

For POM:

Confidence	16	16	64	0.3	0.5	0.2	0.0005	0.001	4	16	0

Passion		16	16	64	0.3	0.5	0.2	0.0005	0.001	4	16	0  

Pleasant	16	16	256	0.1	0.15	0.5	0.0003	0.003	1	128	0

Dominance	8	8	64	0.1	0.2	0.2	0.003	0.0003	4	8	0

Credible	(same model setting as confidence)

Vivid		16	8	64	0.15	0.15	0.15	0.003	0.0005	16	32	0.01

Expertise	8	16	64	0.2	0.15	0.3	0.003	0.0003	16	16	0.001

Entertaining	8	8	64	0.2	0.2	0.15	0.0005	0.001	8	8	0

Reserve		8	16	64	0.5	0.1	0.15	0.003	0.0003	16	8	0.001

Trusting	16	8	128	0.1	0.5	0.15	0.0003	0.001	16	16	0

Relaxed		8	8	64	0.5	0.1	0.2	0.0005	0.0003	8	16	0

Outgoing	16	16	256	0.1	0.15	0.2	0.003	0.0005	4	16	0.01

Thorough	(same model setting as confidence)

Nervous		4	8	256	0	0.5	0.1	0.003	0.003	8	64	0

Sentiment	16	16	64	0.5	0.5	0.2	0.003	0.001	4	8	0.002

Persuasive	(same model setting as confidence)

Humorous	8	8	64	0.5	0.2	0.15	0.003	0.001	4	4	0.001

For IEMOCAP:

Angry		8	4	64	0.3	0.1	0.15	0.003	0.0005	8	64	0.001

Sad		8	4	128	0	0	0	0.0005	0.003	4	64	0.002

Happy		4	16	128	0.3	0.1	0.5	0.003	0.001	1	128	0.002

Neutral		32	8	64	0.2	0.5	0.2	0.003	0.0005	16	16	0.002