Cross-lingual Vision-Language Understanding and Retrieval on IGLUE 1.0 (test) benchmark leaderboard · SOTA2 Research

Cross-lingual Vision-Language Understanding and Retrieval on IGLUE 1.0 (test)

78.95XVNLI AccuracyCCLM_large

Evaluation Results

Method	Links
CCLM_large2022.06		78.95	56.25	74.83	83.78	83.46	43.74	44.88
CCLM base 3M2022.06		75.15	50.94	70.53	66.04	68.15	—	—
CCLM_base2022.06		74.78	48.12	68.49	76.94	76.22	33.9	35.26
CCLM_3M_base2022.06		74.64	42.36	65.91	67.35	65.37	27.46	28.66
VisualBERT2022.06		74.12	48.72	62.35	41.64	36.44	15.36	15.75
VL-BERT2022.06		73.86	49.78	64.16	38.18	31.84	15.11	16.09
UNITER2022.06		73.65	50.62	61.92	41.04	37.49	15.43	16.01
ViLBERT2022.06		73.45	50.33	62.39	36.97	33.21	15.4	16.93
UC22022.06		63.68	42.95	58.32	19.79	17.59	—	—
UC22022.06		62.05	29.35	57.28	20.31	17.89	7.83	9.09
xUNITER2022.06		60.55	40.68	57.46	14.3	13.54	—	—
M3P2022.06		59.36	41.04	49.79	13.21	12.26	—	—
xUNITER2022.06		58.48	21.72	54.59	14.04	13.51	8.72	9.81
M3P2022.06		58.25	28.17	56	12.91	11.9	8.12	9.98
mUNITER2022.06		53.95	37.21	53.41	8.54	9.32	—	—
mUNITER2022.06		53.69	9.97	53.72	8.06	8.86	9.16	10.48