Audio-to-Text Retrieval on AudioCaps

45.1R@1

Proposed

Updated 1mo ago

Evaluation Results

Method
Proposed 2026.04	45.1	79.3	90.5	48.6
Proposed audio-text retrieval framework 2026.04	45.1	79.3	90.5	48.6
LAION-CLAP 2026.04	44.4	76.8	88.9	43.8
Microsoft-CLAP 2026.04	38.1	69.7	81.4	31.9
AudioGPT 2026.04	28.7	44.1	75.1	14.3
Gemini 2026.04	26.6	48.5	76.2	13.6
AudioGPT 2026.04	21.6	37.3	62.3	10.4
Gemini 2026.04	21.4	38.5	42.8	17.2
Gemini 2026.04	20.2	28.9	38.8	15.3
Salmonn 2026.04	19.5	28.7	38.2	20.9
AudioGPT 2026.04	18.3	29.5	57.2	9.3
Salmonn 2026.04	17.2	21.6	30.3	19.8
Salmonn 2026.04	12.5	20.1	22	13.8
LTU 2026.04	11.7	31.5	40.3	11.8
GAMA 2026.04	11.2	29.2	39.4	10.4
LTU 2026.04	10.1	28.9	37.2	9.7
Pengi 2026.04	10	20.7	34.2	10.2
GAMA 2026.04	9.3	18.4	23.8	9.3
LTU 2026.04	9.2	17.3	28.8	8.4
Pengi 2026.04	8.1	18.6	27.5	9.5
GAMA 2026.04	7.3	15.3	20.6	7.2
Pengi 2026.04	5.9	14.7	20.4	3.7