Continual Generalized Zero-Shot Learning on AWA2 benchmark leaderboard · SOTA2 Research

Continual Generalized Zero-Shot Learning on AWA2

89.22Mean Seen Accuracy (mSA)NM-ZSL

Evaluation Results

Method
NM-ZSL2022.03	89.22	51.38	63.41
CV +mem2022.03	88.36	33.24	47.89
A-CZSL2022.03	82.91	42.19	57.74
CA +mem2022.03	82.19	55.98	65.95
CV +mem2022.03	80.97	25.75	38.34
Seq-CADA2022.03	79.89	36.64	47.99
unified feature-generative framework2022.03	79.19	70.71	74.09
DVGR-CZSL2022.03	78.17	40.44	50.89
NM-ZSL2022.03	76.33	39.79	51.51
Seq-CVAE2022.03	73.71	26.22	39.32
DVGR-CZSL2022.03	73.5	28.8	40.6
A-CZSL2022.03	70.16	25.93	37.19
Tf-GCZSL2022.03	67.42	58.08	62.41
unified feature-generative framework2022.03	67.15	54.19	61.32
Tf-GCZSL2022.03	64.89	40.23	48.33
CA +mem2022.03	62.8	39.23	46.22
Seq-CVAE2022.03	61.42	19.34	28.67
AGEM2022.03	57	—	—
EWC2022.03	53.29	—	—
AGEM2022.03	52.61	—	—
Seq-CADA2022.03	52.3	20.3	30.38
EWC2022.03	43.71	—	—
Seq-fCLSWGAN2022.03	22.88	14.92	18.06
Seq-fCLSWGAN2022.03	18.97	8.45	11.69