End-to-end generation on QAMPARI benchmark leaderboard · SOTA2 Research

End-to-end generation on QAMPARI

23.3PrecisionFrank-Wolfe with Exact Line Search

Evaluation Results

Method
Frank-Wolfe with Exact Line Search2026.04	23.3	21.98	19	8.1	12.9
Frank-Wolfe with Exact Line Search2026.04	22.66	22.15	19.06	7.7	12.4
Frank-Wolfe with Exact Line Search2026.04	22.46	19.7	17.68	6.2	11.8
MMR2026.04	22.2	22.24	18.6	6.8	12.3
Frank-Wolfe with Exact Line Search2026.04	22.08	17.5	16.37	4.8	9.9
Top-k2026.04	21.96	21.34	18.24	6	11.4
Top-k2026.04	20.32	24.23	18.47	7.7	12
Frank-Wolfe with Exact Line Search2026.04	20.13	24.53	18.65	8.4	11
Frank-Wolfe with Exact Line Search2026.04	19.9	21.45	17.3	6.2	9.9
Frank-Wolfe with Exact Line Search2026.04	19.88	23.04	17.69	6.6	10.5
MMR2026.04	19.85	21.69	17.38	7.1	10.4
MMR2026.04	19.74	24.77	18.26	8.6	11
Frank-Wolfe with Exact Line Search2026.04	19.52	24.35	18.22	8.2	10.4
DPP2026.04	19.42	21.9	17.26	6.7	10.2
MMR2026.04	19.13	18.99	15.79	5.6	8.8
MMR2026.04	18.49	24.87	17.52	8.6	9.5
DPP2026.04	17.96	16.28	14.17	4	6.7
MMR2026.04	17.77	14.81	13.27	3.4	5.7
DPP2026.04	17.68	23.18	16.52	7.3	9
MMR2026.04	17.59	22.25	16.5	6.7	9.8
DPP2026.04	17.39	7.49	8.42	1.6	3.6
DPP2026.04	16.8	8.38	8.86	1.3	3.9
DPP2026.04	16.72	11.31	11.01	2.1	5.4
MMR2026.04	16.18	16.92	13.56	4.3	6.6
DPP2026.04	15.44	18.46	13.65	4.4	6.3
DPP2026.04	14.86	12.36	10.59	2.5	4.7