Text-to-Speech on English emotional dataset benchmark leaderboard · SOTA2 Research

Text-to-Speech on English emotional dataset

94.8SS ScoreCSP-FT

Evaluation Results

Method	Links
CSP-FT2025.01		94.8	96.8	3.8
Highest Two FT2025.01		94.6	96.5	6.5
Full FT2025.01		94.5	97	12.1
Second Half FT2025.01		93.7	96.6	7.5
Shallowest Two2025.01		93.6	95.5	6.8
Lowest Two FT2025.01		93.6	95.8	4.2
First Half FT2025.01		93.5	95.2	11.5
Deepest Two2025.01		93.5	95	7.2
LoRA FT2025.01		93.3	94.3	8.9
LoRA FT2025.01		93.1	94	8.5
LoRA FT2025.01		93	94.1	9.2
CSP-FT2025.01		92.6	95.5	9.2
Full FT2025.01		92.5	95.6	25.1
Highest Two FT2025.01		92.5	95.3	14.6
Second Half FT2025.01		92.4	95.6	15.7
First Half FT2025.01		92.3	94.2	22.6
Shallowest Two2025.01		92.3	94.7	14.7
Deepest Two2025.01		92.3	94	15.7
Lowest Two FT2025.01		92.3	94.8	12.3
LoRA FT2025.01		92	93.1	18.1
LoRA FT2025.01		91.9	93.6	19.3
LoRA FT2025.01		91.8	93.3	20.6
Origin2025.01		91.2	92	4
Full FT2025.01		90.2	93	9.5
CSP-FT2025.01		90.2	94.2	7
Lowest Two FT2025.01		89.7	93	8
Highest Two FT2025.01		89.7	93.1	10
First Half FT2025.01		89.6	92.9	8.4
LoRA FT2025.01		89.5	91.9	9.9
LoRA FT2025.01		89.5	92.2	9.6
Second Half FT2025.01		89.4	92.3	9.8
Deepest Two2025.01		89.4	92.5	9.8
LoRA FT2025.01		89.2	91.7	9.5
Shallowest Two2025.01		89	92.1	8.4
Origin2025.01		88.9	89.8	8.4
Origin2025.01		84.2	85.6	8.4
CSP-FT2025.01		69.9	83.2	8.5
Full FT2025.01		69.4	83.9	10.3
First Half FT2025.01		69.4	83.2	13.7
Second Half FT2025.01		69.4	82.8	14.5
Highest Two FT2025.01		69.1	83	12.4
Shallowest Two2025.01		68.9	82.5	10.4
LoRA FT2025.01		68.7	82.4	14.7
LoRA FT2025.01		68.7	82.2	12.1
LoRA FT2025.01		68.7	82.4	11.8
Deepest Two2025.01		68.5	82.4	13.7
Lowest Two FT2025.01		68.3	82.2	9.5
Origin2025.01		64	77.7	8