Share your thoughts, 1 month free Claude Pro on usSee more

Data-to-Text Generation on DART (test)

52BLEU

C-P (large)

Updated 2d ago

Evaluation Results

Method
C-P (large) 2023.08	52	-	-	-	-	-	-	-	-
T5-3B 2023.08	50.2	-	-	-	-	-	-	-	-
T5-base 2023.08	49.8	-	-	-	-	-	-	-	-
T5-3B 2023.08	49.8	-	-	-	-	-	-	-	-
T5-3B 2023.08	49.6	-	-	-	-	-	-	-	-
T5-3B 2023.08	49.3	-	-	-	-	-	-	-	-
T5-base 2024.01	49.21	40	44	-	-	-	-	-	-
T5-base 2023.08	49	-	-	-	-	-	-	-	-
T5-base 2023.08	48.6	-	-	-	-	-	-	-	-
T5-base 2024.01	48.52	39.7	45.91	-	-	-	-	-	-
T5-base 2023.08	48.1	-	-	-	-	-	-	-	-
self-mem + new data 2024.01	47.76	39.51	48.28	-	-	-	-	-	-
self-mem + new data + self T2D 2024.01	47.6	39.39	48.23	-	-	-	-	-	-
self-mem + new data 2024.01	47.54	39.38	48.07	-	-	-	-	-	-
LoRA 2021.06	47.5	39	45	-	-	-	-	-	-
self-mem + new data + self T2D 2024.01	47.32	39.19	47.34	-	-	-	-	-	-
BART-base 2024.01	47.11	38	46	-	-	-	-	-	-
LoRA 2021.06	47.1	39	46	-	-	-	-	-	-
Adapter (H) 2021.06	47.1	39	45	-	-	-	-	-	-
Fine-Tune 2021.06	47	39	46	-	-	-	-	-	-
no self-mem 3 2024.01	46.9	38.92	46.61	-	-	-	-	-	-
PrefLayer 2021.06	46.7	38	45	-	-	-	-	-	-
UnifiedSKG (3B) 2023.08	46.7	-	-	-	-	-	-	-	-
PrefLayer 2021.06	46.4	38	46	-	-	-	-	-	-
Fine-Tune 2021.06	46.2	39	46	-	-	-	-	-	-
UnifiedSKG (base) 2023.08	46.2	-	-	-	-	-	-	-	-
BART-base 2024.01	46.15	38.13	48.2	-	-	-	-	-	-
self-mem + self T2D 2024.01	46.14	38.86	49.08	-	-	-	-	-	-
Adapter (L) 2021.06	45.7	38	46	-	-	-	-	-	-
self-mem 2024.01	45.63	38.9	49.1	-	-	-	-	-	-
self-mem + self T2D 2024.01	45.62	38.81	48.95	-	-	-	-	-	-
FLAN-T5-base 2024.01	45.53	37.36	48.54	-	-	-	-	-	-
AdaPreLoRA AdamW 2026.05	45.4	67	0.49	-	60	40	-	-	-
LoRA 2025.06	45.35	38	53	-	-	-	-	-	-
Adapter (H) 2021.06	45.2	38	46	-	-	-	-	-	-
LoRA-Pro AdamW 2026.05	44.9	66	0.5	-	62	39	-	-	-
no self-mem 1 2024.01	44.86	37.98	50.83	-	-	-	-	-	-
Scaled AdamW 2026.05	44.8	67	0.49	-	62	40	-	-	-
AdaPreLoRA SGD 2026.05	44.6	66	0.49	-	62	39	-	-	-
self-mem 2024.01	44.52	38.62	50.75	-	-	-	-	-	-
LoRA-Pro SGD 2026.05	44.1	66	0.5	-	61	38	-	-	-
AdamW 2026.05	43.9	66	0.5	-	60	38	-	-	-
Scaled GD 2026.05	43.8	66	0.5	-	61	38	-	-	-
LoRMA+ 2025.06	43.64	38	53	-	-	-	-	-	-
Adapter (L) 2021.06	42.4	36	48	-	-	-	-	-	-
no self-mem 2 2024.01	41.23	33.98	51.47	-	-	-	-	-	-
SGD 2026.05	41.2	63	0.52	-	59	33	-	-	-
FT (Top2) 2021.06	41	34	56	-	-	-	-	-	-
SVD Denoising 2025.10	33.29	57.9	-	51.95	-	-	-	5.26	1.26
LSTM with Attention 2024.01	29.66	27	63	-	-	-	-	-	-
End-to-End Transformer 2024.01	27.24	25	65	-	-	-	-	-	-
SVD Denoising 2025.10	23.58	46.2	-	46.46	-	-	-	2.87	0.84
SVD Denoising 2025.10	21.78	48.4	-	46.98	-	-	-	4.22	0.86
DP-LoRA 2025.10	21.77	48.3	-	46.98	-	-	-	4.23	0.86
DP-LoRA 2025.10	21.74	52.2	-	44.47	-	-	-	3.84	0.86
DP-LoRA 2025.10	14.66	32.3	-	33.44	-	-	-	0.91	0.59
SVD Denoising 2025.10	13.81	45.4	-	43.12	-	-	-	3.68	0.73
SVD Denoising 2025.10	9.47	37.5	-	37.33	-	-	-	2.95	0.47
DP-LoRA 2025.10	8.88	36.8	-	37.38	-	-	-	2.86	0.41
DP-MuonBC 2026.05	7.5	-	-	34.42	-	-	0.446	-	-
DP-Adam 2026.05	7.46	-	-	34.39	-	-	0.4697	-	-
DP-Muon 2026.05	7.41	-	-	34.38	-	-	0.4531	-	-
DP-SGD 2026.05	6.12	-	-	33.19	-	-	0.6287	-	-
DP-LoRA 2025.10	6.12	31.9	-	30.08	-	-	-	2.13	0.26
OPT 2022.12	-	-	-	14.4	-	-	-	-	-
OPT 2022.12	-	-	-	40.6	-	-	-	-	-
OPT-IML 2022.12	-	-	-	43	-	-	-	-	-
OPT-IML 2022.12	-	-	-	44.3	-	-	-	-	-
OPT 2022.12	-	-	-	22.5	-	-	-	-	-
OPT 2022.12	-	-	-	48.7	-	-	-	-	-
OPT-IML 2022.12	-	-	-	44.1	-	-	-	-	-
OPT-IML 2022.12	-	-	-	49.8	-	-	-	-	-