General Language Understanding on GLUE v1 (test dev)

87.86MNLIBOMF (Ours)

Evaluation Results

Method
BOMF (Ours)2024.11	87.86	—	93.5	95.54	—	—	81.4	—	—	93.9	88.68
BOMF2024.11	87.86	—	93.5	95.54	—	—	81.4	—	90.15	93.9	88.68
Greedy SWA2024.11	87.84	—	93.3	95.2	—	—	80.01	—	89.73	93.38	88.67
Learned SWA2024.11	87.55	—	93.03	95.2	—	—	78.79	—	89.43	93.56	88.43
Grid Search Inner BO2024.11	87.54	—	92.79	95.31	—	—	79.66	—	89.33	92.48	88.24
BOMFT (Ours)2024.11	87.51	—	94.83	95.65	—	—	81.75	—	—	93.37	88.66
BOMF2024.11	87.51	—	94.83	95.65	—	—	81.75	—	90.3	93.37	88.66
HPBO (Full)2024.11	87.46	—	93.01	95.11	—	—	78.57	—	—	92.78	88.58
HPBO2024.11	87.46	—	93.01	95.11	—	—	78.57	—	89.25	92.78	88.58
Greedy SWA2024.11	87.45	—	93.16	95.54	—	—	80.7	—	—	92.83	88.64
Greedy SWA2024.11	87.45	—	93.16	95.54	—	—	80.7	—	89.72	92.83	88.64
HPBO2024.11	87.45	—	92.71	94.41	—	—	78.49	—	88.97	92.72	88.04
SWA2024.11	87.44	—	93.21	94.75	—	—	76.7	—	88.7	91.73	88.35
TWA2024.11	87.44	—	92.86	94.87	—	—	78.29	—	89.03	92.16	88.59
OTFusion2024.11	87.43	—	92.22	94.27	—	—	77.08	—	—	92.82	88.34
TWA2024.11	87.42	—	93	95.54	—	—	81.23	—	—	91.58	87.85
TWA2024.11	87.42	—	93	95.54	—	—	81.23	—	89.44	91.58	87.85
Grid Fine-Tune2024.11	87.41	—	92.62	94.87	—	—	77.78	—	—	92.39	88.16
SWA2024.11	87.41	—	92.81	95.42	—	—	78.62	—	—	92.24	88.49
Learned SWA2024.11	87.41	—	92.94	95.31	—	—	81.4	—	—	92.81	88.38
Grid Fine-tune2024.11	87.41	—	92.62	94.87	—	—	77.78	—	88.93	92.39	88.16
SWA2024.11	87.41	—	92.81	95.42	—	—	78.62	—	89.17	92.24	88.49
Learned SWA2024.11	87.41	—	92.94	95.31	—	—	81.4	—	89.71	92.81	88.38
SCALEARN++2023.10	87.06	90.04	92.03	94.15	90.62	88.21	80.87	59.82	85.35	—	—
SCALEARNUNIFORM++2023.10	86.98	90.38	92.53	94.11	90.18	87.43	80.04	59.45	85.14	—	—
SCALEARN2023.10	86.97	90.32	92.51	93.88	90.96	87.75	82.06	58.47	85.36	—	—
SCALEARNUNIFORM2023.10	86.93	90.38	92.53	93.58	90.08	87.57	80.07	59.04	85.02	—	—
ADAPTERFUSION2023.10	86.82	90.23	92.48	93.23	90.37	88.41	79.49	59.04	85.01	—	—
FINETUNE2023.10	86.61	90.32	91.78	93.33	90.53	86.94	73.47	58.46	83.93	—	—
LORA2023.10	86.52	89.86	92.25	94.19	90.66	87.03	70.4	57.55	83.56	—	—
ADAPTER2023.10	86.5	90.18	92.25	93.65	90.23	86.64	72.89	58.28	83.83	—	—
HYPERFORMER++2023.10	86.38	88.81	91.99	93.27	90.8	87.83	83.75	54.05	84.61	—	—
PROPETL2023.10	86.19	88.88	92.05	93.81	90.03	85.93	74.19	59.29	83.8	—	—
HYPERFORMER2023.10	86.08	89.13	91.81	93.16	90.63	87.01	82.79	57.3	84.74	—	—
ADAPTER-M2023.10	86.03	89.69	91.58	93.35	88.71	86.76	80.26	51.79	83.52	—	—
COMPACTER++2023.10	85.62	88.84	91.79	93.58	89.67	87.21	72.02	58.49	83.4	—	—
PROPETL-M2023.10	85.23	87.82	91.37	93.88	90.27	86.36	78.58	54.71	83.53	—	—
FINETUNE-M2023.10	84.95	89.76	90.91	92.58	86.14	83.42	80.99	49.12	82.23	—	—
(IA)32023.10	83.78	88.37	90.57	93.35	89.93	87.11	72.56	56.57	82.78	—	—
ADAPTERSOUP2023.10	63.47	81.63	78	90.75	80.17	75	62.09	41.06	71.52	—	—