binwang commited on
Commit
8652063
·
verified ·
1 Parent(s): 02ce699

Upload organize_model_results.json with huggingface_hub

Browse files
Files changed (1) hide show
  1. organize_model_results.json +121 -41
organize_model_results.json CHANGED
@@ -5,6 +5,7 @@
5
  "MERaLiON-AudioLLM-Whisper-SEA-LION": 47.01682396389003,
6
  "Qwen2-Audio-7B-Instruct": 29.187525646286417,
7
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 24.640951990151827,
 
8
  "WavLLM_fairseq": 39.96717275338531,
9
  "SALMONN_7B": 34.222404595814524,
10
  "cascade_whisper_large_v3_llama_3_8b_instruct": 39.32704144439885
@@ -13,12 +14,22 @@
13
  "cascade_whisper_large_v3_llama_3_8b_instruct": 39.462453836684446
14
  }
15
  },
 
 
 
 
 
 
 
 
 
16
  "wavcaps_test": {
17
  "meteor": {
18
  "Qwen-Audio-Chat": 0.2355106805560457,
19
  "MERaLiON-AudioLLM-Whisper-SEA-LION": 0.3175511907248581,
20
  "Qwen2-Audio-7B-Instruct": 0.21342294856199182,
21
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 0.120421856260385,
 
22
  "WavLLM_fairseq": 0.06399522524688675,
23
  "SALMONN_7B": 0.17175112770658157,
24
  "cascade_whisper_large_v3_llama_3_8b_instruct": 0.1388630786594543
@@ -28,6 +39,7 @@
28
  "MERaLiON-AudioLLM-Whisper-SEA-LION": 33.97687861271676,
29
  "Qwen2-Audio-7B-Instruct": 33.78034682080925,
30
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 6.3468208092485545,
 
31
  "WavLLM_fairseq": 6.901734104046243,
32
  "SALMONN_7B": 23.76878612716763,
33
  "cascade_whisper_large_v3_llama_3_8b_instruct": 3.445086705202312
@@ -43,6 +55,7 @@
43
  "Qwen2-Audio-7B-Instruct": 16.466557744958333,
44
  "whisper_large_v3": 14.673689493155793,
45
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 15.209998552437538,
 
46
  "WavLLM_fairseq": 2.368659001743569,
47
  "SALMONN_7B": 5.296039450108202,
48
  "cascade_whisper_large_v3_llama_3_8b_instruct": 14.154700735606419
@@ -54,6 +67,7 @@
54
  "MERaLiON-AudioLLM-Whisper-SEA-LION": 67.2,
55
  "Qwen2-Audio-7B-Instruct": 53.6,
56
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 71.6,
 
57
  "WavLLM_fairseq": 62.199999999999996,
58
  "SALMONN_7B": 46.8,
59
  "cascade_whisper_large_v3_llama_3_8b_instruct": 64.0
@@ -92,6 +106,7 @@
92
  "MERaLiON-AudioLLM-Whisper-SEA-LION": 57.7927548441449,
93
  "Qwen2-Audio-7B-Instruct": 71.60909856781802,
94
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 51.727042965459134,
 
95
  "WavLLM_fairseq": 44.3133951137321,
96
  "SALMONN_7B": 50.88458298230834,
97
  "cascade_whisper_large_v3_llama_3_8b_instruct": 56.44481887110362
@@ -116,6 +131,7 @@
116
  "Qwen2-Audio-7B-Instruct": 0.1905689473257041,
117
  "whisper_large_v3": 0.3171008846684522,
118
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 0.32988393799204613,
 
119
  "WavLLM_fairseq": 0.4463923382842302,
120
  "SALMONN_7B": 0.42346400454508565,
121
  "cascade_whisper_large_v3_llama_3_8b_instruct": 0.31912994075156237
@@ -128,6 +144,7 @@
128
  "Qwen2-Audio-7B-Instruct": 0.18872219319407232,
129
  "whisper_large_v3": 0.11863959266711877,
130
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 0.11416493424197618,
 
131
  "WavLLM_fairseq": 0.6447482518259942,
132
  "SALMONN_7B": 0.2577708974886327,
133
  "cascade_whisper_large_v3_llama_3_8b_instruct": 0.11773910240019567
@@ -152,6 +169,7 @@
152
  "Qwen2-Audio-7B-Instruct": 0.060415760304159495,
153
  "whisper_large_v3": 0.03660128246354058,
154
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 0.05307658841999735,
 
155
  "WavLLM_fairseq": 0.04798834811886432,
156
  "SALMONN_7B": 0.09671439650443565,
157
  "cascade_whisper_large_v3_llama_3_8b_instruct": 0.03714982881570734
@@ -164,6 +182,7 @@
164
  "Qwen2-Audio-7B-Instruct": 0.035141660693401744,
165
  "whisper_large_v3": 0.01878749009695552,
166
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 0.032349945297468596,
 
167
  "WavLLM_fairseq": 0.02103218017882069,
168
  "SALMONN_7B": 0.10270871845172973,
169
  "cascade_whisper_large_v3_llama_3_8b_instruct": 0.018334779492209605
@@ -176,6 +195,7 @@
176
  "Qwen2-Audio-7B-Instruct": 0.2245352799625317,
177
  "whisper_large_v3": 0.1698509342851144,
178
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 0.1789273082575623,
 
179
  "WavLLM_fairseq": 0.42541061709652933,
180
  "SALMONN_7B": 0.24872817713464365,
181
  "cascade_whisper_large_v3_llama_3_8b_instruct": 0.17467982364056267
@@ -187,6 +207,7 @@
187
  "MERaLiON-AudioLLM-Whisper-SEA-LION": 65.6,
188
  "Qwen2-Audio-7B-Instruct": 44.800000000000004,
189
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 72.2,
 
190
  "WavLLM_fairseq": 19.2,
191
  "SALMONN_7B": 15.8,
192
  "cascade_whisper_large_v3_llama_3_8b_instruct": 63.0
@@ -201,6 +222,7 @@
201
  "MERaLiON-AudioLLM-Whisper-SEA-LION": 48.505976095617534,
202
  "Qwen2-Audio-7B-Instruct": 53.98406374501992,
203
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 44.322709163346616,
 
204
  "WavLLM_fairseq": 59.76095617529881,
205
  "SALMONN_7B": 23.804780876494025,
206
  "cascade_whisper_large_v3_llama_3_8b_instruct": 46.713147410358566
@@ -212,6 +234,7 @@
212
  "MERaLiON-AudioLLM-Whisper-SEA-LION": 59.7093023255814,
213
  "Qwen2-Audio-7B-Instruct": 58.31395348837209,
214
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 73.11046511627907,
 
215
  "WavLLM_fairseq": 58.54651162790698,
216
  "SALMONN_7B": 59.24418604651163,
217
  "cascade_whisper_large_v3_llama_3_8b_instruct": 64.94186046511628
@@ -239,6 +262,7 @@
239
  "Qwen2-Audio-7B-Instruct": 0.11438872500819404,
240
  "whisper_large_v3": 0.10001863741235596,
241
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 0.10600831614192711,
 
242
  "WavLLM_fairseq": 0.14533325621300636,
243
  "SALMONN_7B": 0.3062255383962828,
244
  "cascade_whisper_large_v3_llama_3_8b_instruct": 0.09876543209876543
@@ -250,6 +274,7 @@
250
  "MERaLiON-AudioLLM-Whisper-SEA-LION": 73.66473556344609,
251
  "Qwen2-Audio-7B-Instruct": 64.86264249672958,
252
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 88.61894972902262,
 
253
  "WavLLM_fairseq": 77.64903756307233,
254
  "SALMONN_7B": 66.39506634273968,
255
  "cascade_whisper_large_v3_llama_3_8b_instruct": 83.81984675761541
@@ -266,6 +291,7 @@
266
  "whisper_large_v3": 0.5377268970583734,
267
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 0.5840399155162387,
268
  "gemini-1.5-flash": 1.1100431601824359,
 
269
  "WavLLM_fairseq": 1.2204842511249197,
270
  "SALMONN_7B": 1.0189782362484312,
271
  "cascade_whisper_large_v3_llama_3_8b_instruct": 0.507882090054792
@@ -277,6 +303,7 @@
277
  "MERaLiON-AudioLLM-Whisper-SEA-LION": 46.206896551724135,
278
  "Qwen2-Audio-7B-Instruct": 53.9463601532567,
279
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 56.59003831417625,
 
280
  "WavLLM_fairseq": 51.072796934865906,
281
  "SALMONN_7B": 41.7624521072797,
282
  "cascade_whisper_large_v3_llama_3_8b_instruct": 45.593869731800766
@@ -288,6 +315,7 @@
288
  "MERaLiON-AudioLLM-Whisper-SEA-LION": 53.2,
289
  "Qwen2-Audio-7B-Instruct": 39.6,
290
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 66.0,
 
291
  "WavLLM_fairseq": 46.6,
292
  "SALMONN_7B": 36.6,
293
  "cascade_whisper_large_v3_llama_3_8b_instruct": 53.8
@@ -302,6 +330,7 @@
302
  "MERaLiON-AudioLLM-Whisper-SEA-LION": 99.75379565038982,
303
  "Qwen2-Audio-7B-Instruct": 99.1177677472302,
304
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 34.94050061551087,
 
305
  "WavLLM_fairseq": 69.61427985227739,
306
  "SALMONN_7B": 88.79770209273697,
307
  "cascade_whisper_large_v3_llama_3_8b_instruct": 42.921624948707425
@@ -313,6 +342,7 @@
313
  "MERaLiON-AudioLLM-Whisper-SEA-LION": 93.76666666666667,
314
  "Qwen2-Audio-7B-Instruct": 61.56666666666667,
315
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 19.6,
 
316
  "WavLLM_fairseq": 46.766666666666666,
317
  "SALMONN_7B": 42.733333333333334,
318
  "cascade_whisper_large_v3_llama_3_8b_instruct": 25.433333333333337
@@ -337,6 +367,7 @@
337
  "Qwen2-Audio-7B-Instruct": 0.27856006770658537,
338
  "whisper_large_v3": 0.2143555471246589,
339
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 0.22881615619208825,
 
340
  "WavLLM_fairseq": 0.39796588405247263,
341
  "SALMONN_7B": 0.34868891450584405,
342
  "cascade_whisper_large_v3_llama_3_8b_instruct": 0.22004640235805695
@@ -360,6 +391,7 @@
360
  "Qwen2-Audio-7B-Instruct": 0.23542555661330924,
361
  "whisper_large_v3": 0.15887899737116104,
362
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 0.1448629161356777,
 
363
  "WavLLM_fairseq": 0.6671766188447099,
364
  "SALMONN_7B": 0.3597423676988383,
365
  "cascade_whisper_large_v3_llama_3_8b_instruct": 0.15611126487402763
@@ -396,6 +428,7 @@
396
  "Qwen2-Audio-7B-Instruct": 6.326113431899141,
397
  "whisper_large_v3": 46.01512198258627,
398
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 46.80524126004861,
 
399
  "WavLLM_fairseq": 5.933522277713613,
400
  "SALMONN_7B": 26.89649039333571,
401
  "cascade_whisper_large_v3_llama_3_8b_instruct": 46.79924664837527
@@ -432,6 +465,7 @@
432
  "Qwen2-Audio-7B-Instruct": 0.11723812890302816,
433
  "whisper_large_v3": 0.09459022434812692,
434
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 0.09948381629977261,
 
435
  "WavLLM_fairseq": 0.15491778414546403,
436
  "SALMONN_7B": 0.10765150204693537,
437
  "cascade_whisper_large_v3_llama_3_8b_instruct": 0.09515429104337297
@@ -455,6 +489,7 @@
455
  "MERaLiON-AudioLLM-Whisper-SEA-LION": 48.4,
456
  "Qwen2-Audio-7B-Instruct": 33.8,
457
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 45.4,
 
458
  "WavLLM_fairseq": 31.6,
459
  "SALMONN_7B": 9.0,
460
  "cascade_whisper_large_v3_llama_3_8b_instruct": 37.400000000000006
@@ -469,6 +504,7 @@
469
  "MERaLiON-AudioLLM-Whisper-SEA-LION": 77.83333333333333,
470
  "Qwen2-Audio-7B-Instruct": 0.9666666666666667,
471
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 7.633333333333334,
 
472
  "WavLLM_fairseq": 0.23333333333333336,
473
  "SALMONN_7B": 0.06666666666666667,
474
  "cascade_whisper_large_v3_llama_3_8b_instruct": 9.666666666666666
@@ -480,6 +516,7 @@
480
  "MERaLiON-AudioLLM-Whisper-SEA-LION": 93.48605577689243,
481
  "Qwen2-Audio-7B-Instruct": 92.80876494023903,
482
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 15.737051792828685,
 
483
  "WavLLM_fairseq": 51.932270916334666,
484
  "SALMONN_7B": 81.31474103585658,
485
  "cascade_whisper_large_v3_llama_3_8b_instruct": 44.22310756972111
@@ -492,6 +529,7 @@
492
  "Qwen2-Audio-7B-Instruct": 0.2080008649583739,
493
  "whisper_large_v3": 0.17210509244242622,
494
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 0.2192622950819672,
 
495
  "WavLLM_fairseq": 0.48091685587631094,
496
  "SALMONN_7B": 0.3238620391393664,
497
  "cascade_whisper_large_v3_llama_3_8b_instruct": 0.23561466104443723
@@ -516,6 +554,7 @@
516
  "Qwen2-Audio-7B-Instruct": 74.7247908410392,
517
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 91.85380889476001,
518
  "gemini-1.5-flash": 89.25583443416997,
 
519
  "WavLLM_fairseq": 66.31439894319684,
520
  "SALMONN_7B": 50.99075297225891,
521
  "cascade_whisper_large_v3_llama_3_8b_instruct": 85.2928225451343
@@ -527,6 +566,7 @@
527
  "MERaLiON-AudioLLM-Whisper-SEA-LION": 84.31782540512285,
528
  "Qwen2-Audio-7B-Instruct": 66.49242028227914,
529
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 89.33612127548353,
 
530
  "WavLLM_fairseq": 66.5446941975954,
531
  "SALMONN_7B": 56.455828541557764,
532
  "cascade_whisper_large_v3_llama_3_8b_instruct": 86.4610559330894
@@ -538,6 +578,7 @@
538
  "MERaLiON-AudioLLM-Whisper-SEA-LION": 57.0,
539
  "Qwen2-Audio-7B-Instruct": 40.4,
540
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 58.0,
 
541
  "WavLLM_fairseq": 45.199999999999996,
542
  "SALMONN_7B": 17.2,
543
  "cascade_whisper_large_v3_llama_3_8b_instruct": 49.0
@@ -553,6 +594,7 @@
553
  "Qwen2-Audio-7B-Instruct": 0.09260359129694522,
554
  "whisper_large_v3": 0.12359684029221357,
555
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 0.20886539565639167,
 
556
  "WavLLM_fairseq": 0.7054601967888183,
557
  "SALMONN_7B": 0.8259290055631446,
558
  "cascade_whisper_large_v3_llama_3_8b_instruct": 0.12450753301261111
@@ -564,6 +606,7 @@
564
  "MERaLiON-AudioLLM-Whisper-SEA-LION": 51.4,
565
  "Qwen2-Audio-7B-Instruct": 42.0,
566
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 56.0,
 
567
  "WavLLM_fairseq": 45.199999999999996,
568
  "SALMONN_7B": 40.599999999999994,
569
  "cascade_whisper_large_v3_llama_3_8b_instruct": 49.0
@@ -578,6 +621,7 @@
578
  "MERaLiON-AudioLLM-Whisper-SEA-LION": 46.4,
579
  "Qwen2-Audio-7B-Instruct": 24.8,
580
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 44.0,
 
581
  "WavLLM_fairseq": 31.6,
582
  "SALMONN_7B": 7.0,
583
  "cascade_whisper_large_v3_llama_3_8b_instruct": 36.0
@@ -604,6 +648,7 @@
604
  "MERaLiON-AudioLLM-Whisper-SEA-LION": 38.00454545454545,
605
  "Qwen2-Audio-7B-Instruct": 40.77727272727273,
606
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 3.0954545454545457,
 
607
  "WavLLM_fairseq": 5.5,
608
  "SALMONN_7B": 37.445454545454545,
609
  "cascade_whisper_large_v3_llama_3_8b_instruct": 2.4727272727272727
@@ -613,6 +658,7 @@
613
  "MERaLiON-AudioLLM-Whisper-SEA-LION": 0.24920047034353812,
614
  "Qwen2-Audio-7B-Instruct": 0.19891712076314283,
615
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 0.05796819723943051,
 
616
  "WavLLM_fairseq": 0.041732965094428545,
617
  "SALMONN_7B": 0.20994052484339956,
618
  "cascade_whisper_large_v3_llama_3_8b_instruct": 0.07953048457785493
@@ -627,6 +673,7 @@
627
  "MERaLiON-AudioLLM-Whisper-SEA-LION": 7.816666666666666,
628
  "Qwen2-Audio-7B-Instruct": 2.55,
629
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 26.016666666666666,
 
630
  "WavLLM_fairseq": 2.6833333333333336,
631
  "SALMONN_7B": 2.5166666666666666,
632
  "cascade_whisper_large_v3_llama_3_8b_instruct": 12.416666666666666
@@ -651,6 +698,7 @@
651
  "Qwen2-Audio-7B-Instruct": 0.04425838146050298,
652
  "whisper_large_v3": 2.451098639578599,
653
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 2.8327095799289337,
 
654
  "WavLLM_fairseq": 0.1695522548322915,
655
  "SALMONN_7B": 0.3649023706010388,
656
  "cascade_whisper_large_v3_llama_3_8b_instruct": 2.4245628096245917
@@ -663,6 +711,7 @@
663
  "Qwen2-Audio-7B-Instruct": 16.325186897428104,
664
  "whisper_large_v3": 1.600581653970121,
665
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 27.620150160643625,
 
666
  "WavLLM_fairseq": 13.841886973016162,
667
  "SALMONN_7B": 14.102682915273142,
668
  "cascade_whisper_large_v3_llama_3_8b_instruct": 10.930203684508578
@@ -674,6 +723,7 @@
674
  "MERaLiON-AudioLLM-Whisper-SEA-LION": 63.15021876519203,
675
  "Qwen2-Audio-7B-Instruct": 50.919591292758774,
676
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 24.647544968400585,
 
677
  "WavLLM_fairseq": 43.01199466903598,
678
  "SALMONN_7B": 57.75401069518716,
679
  "cascade_whisper_large_v3_llama_3_8b_instruct": 29.47134606841404
@@ -690,6 +740,7 @@
690
  "whisper_large_v3": 0.12226319428439733,
691
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 0.1400092187139894,
692
  "gemini-1.5-flash": 0.1089344703080587,
 
693
  "WavLLM_fairseq": 0.41876008296842593,
694
  "SALMONN_7B": 0.21487285856956287,
695
  "cascade_whisper_large_v3_llama_3_8b_instruct": 0.12579703464700007
@@ -702,6 +753,7 @@
702
  "Qwen2-Audio-7B-Instruct": 0.35076166942732234,
703
  "whisper_large_v3": 0.27026366524560785,
704
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 0.3035544573275043,
 
705
  "WavLLM_fairseq": 0.7540934640345399,
706
  "SALMONN_7B": 0.6569229098215983,
707
  "cascade_whisper_large_v3_llama_3_8b_instruct": 0.29992939962527493
@@ -713,6 +765,7 @@
713
  "MERaLiON-AudioLLM-Whisper-SEA-LION": 74.80000000000001,
714
  "Qwen2-Audio-7B-Instruct": 52.599999999999994,
715
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 73.8,
 
716
  "WavLLM_fairseq": 21.6,
717
  "SALMONN_7B": 17.2,
718
  "cascade_whisper_large_v3_llama_3_8b_instruct": 70.8
@@ -750,6 +803,7 @@
750
  "MERaLiON-AudioLLM-Whisper-SEA-LION": 49.77635782747604,
751
  "Qwen2-Audio-7B-Instruct": 45.75079872204473,
752
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 18.466453674121407,
 
753
  "WavLLM_fairseq": 29.840255591054312,
754
  "SALMONN_7B": 50.287539936102235,
755
  "cascade_whisper_large_v3_llama_3_8b_instruct": 17.380191693290733
@@ -765,6 +819,7 @@
765
  "Qwen2-Audio-7B-Instruct": 0.07197717796796138,
766
  "whisper_large_v3": 0.06844171360300393,
767
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 0.07041669714480775,
 
768
  "WavLLM_fairseq": 0.10077292565771828,
769
  "SALMONN_7B": 0.0925804013361617,
770
  "cascade_whisper_large_v3_llama_3_8b_instruct": 0.06922195401458074
@@ -789,6 +844,7 @@
789
  "Qwen2-Audio-7B-Instruct": 0.03245972071872916,
790
  "whisper_large_v3": 0.02107778621423822,
791
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 8.433062902024755,
 
792
  "WavLLM_fairseq": 0.0033159224040994286,
793
  "SALMONN_7B": 0.00046745670226766583,
794
  "cascade_whisper_large_v3_llama_3_8b_instruct": 1.0368044741318085
@@ -800,6 +856,7 @@
800
  "MERaLiON-AudioLLM-Whisper-SEA-LION": 46.31578947368421,
801
  "Qwen2-Audio-7B-Instruct": 44.473684210526315,
802
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 18.88157894736842,
 
803
  "WavLLM_fairseq": 26.25,
804
  "SALMONN_7B": 47.30263157894737,
805
  "cascade_whisper_large_v3_llama_3_8b_instruct": 16.710526315789473
@@ -827,11 +884,68 @@
827
  "Qwen2-Audio-7B-Instruct": 25.765420247070075,
828
  "whisper_large_v3": 0.16408986541757878,
829
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 35.274306071307024,
 
830
  "WavLLM_fairseq": 31.96381187282953,
831
  "SALMONN_7B": 33.88941292215531,
832
  "cascade_whisper_large_v3_llama_3_8b_instruct": 5.987143868370054
833
  }
834
  },
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
835
  "imda_part5_30s_sqa_test": {
836
  "llama3_70b_judge": {
837
  "Qwen-Audio-Chat": 61.260000000000005,
@@ -850,6 +964,7 @@
850
  "MERaLiON-AudioLLM-Whisper-SEA-LION": 66.13333333333333,
851
  "Qwen2-Audio-7B-Instruct": 68.38333333333333,
852
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 26.35,
 
853
  "WavLLM_fairseq": 49.06666666666666,
854
  "SALMONN_7B": 59.766666666666666,
855
  "cascade_whisper_large_v3_llama_3_8b_instruct": 36.016666666666666
@@ -861,6 +976,7 @@
861
  "MERaLiON-AudioLLM-Whisper-SEA-LION": 86.76470588235293,
862
  "Qwen2-Audio-7B-Instruct": 80.04901960784315,
863
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 88.57843137254902,
 
864
  "WavLLM_fairseq": 83.92156862745098,
865
  "SALMONN_7B": 83.48039215686273,
866
  "cascade_whisper_large_v3_llama_3_8b_instruct": 82.99019607843137
@@ -876,6 +992,7 @@
876
  "Qwen2-Audio-7B-Instruct": 0.08739585179932637,
877
  "whisper_large_v3": 0.03208650948413402,
878
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 0.04396383619925545,
 
879
  "WavLLM_fairseq": 0.4536784258110264,
880
  "SALMONN_7B": 0.14231519234178336,
881
  "cascade_whisper_large_v3_llama_3_8b_instruct": 0.04754476156709803
@@ -887,6 +1004,7 @@
887
  "MERaLiON-AudioLLM-Whisper-SEA-LION": 36.36015325670498,
888
  "Qwen2-Audio-7B-Instruct": 41.60919540229885,
889
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 47.356321839080465,
 
890
  "WavLLM_fairseq": 41.57088122605364,
891
  "SALMONN_7B": 30.536398467432953,
892
  "cascade_whisper_large_v3_llama_3_8b_instruct": 36.81992337164751
@@ -899,6 +1017,7 @@
899
  "Qwen2-Audio-7B-Instruct": 0.06114048472375004,
900
  "whisper_large_v3": 0.037649480146197796,
901
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 0.04900464852205386,
 
902
  "WavLLM_fairseq": 0.06621482559171073,
903
  "SALMONN_7B": 0.0459884319222171,
904
  "cascade_whisper_large_v3_llama_3_8b_instruct": 0.038146268762641496
@@ -912,6 +1031,7 @@
912
  "whisper_large_v3": 0.7225930420711975,
913
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 0.7824973031283711,
914
  "gemini-1.5-flash": 0.9690871089536138,
 
915
  "WavLLM_fairseq": 1.2913969795037756,
916
  "SALMONN_7B": 1.2721817691477886,
917
  "cascade_whisper_large_v3_llama_3_8b_instruct": 0.6848705501618123
@@ -935,6 +1055,7 @@
935
  "MERaLiON-AudioLLM-Whisper-SEA-LION": 64.80000000000001,
936
  "Qwen2-Audio-7B-Instruct": 51.6,
937
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 74.0,
 
938
  "WavLLM_fairseq": 50.8,
939
  "SALMONN_7B": 44.6,
940
  "cascade_whisper_large_v3_llama_3_8b_instruct": 57.800000000000004
@@ -1006,47 +1127,6 @@
1006
  "cascade_whisper_large_v3_llama_3_8b_instruct": 63.68000000000001
1007
  }
1008
  },
1009
- "imda_part4_30s_asr_test": {
1010
- "wer": {
1011
- "MERaLiON-AudioLLM-Whisper-SEA-LION": 0.0
1012
- }
1013
- },
1014
- "mmau_mini": {
1015
- "string_match": {
1016
- "MERaLiON-AudioLLM-Whisper-SEA-LION": 60.5
1017
- },
1018
- "llama3_70b_judge": {
1019
- "MERaLiON-AudioLLM-Whisper-SEA-LION": 64.60000000000001,
1020
- "phi_4_multimodal_instruct": 59.4
1021
- }
1022
- },
1023
- "mmau_mini_music": {
1024
- "string_match": {
1025
- "MERaLiON-AudioLLM-Whisper-SEA-LION": 0.6077844311377245
1026
- },
1027
- "llama3_70b_judge": {
1028
- "MERaLiON-AudioLLM-Whisper-SEA-LION": 0.6437125748502994,
1029
- "phi_4_multimodal_instruct": 0.688622754491018
1030
- }
1031
- },
1032
- "mmau_mini_sound": {
1033
- "string_match": {
1034
- "MERaLiON-AudioLLM-Whisper-SEA-LION": 0.6606606606606606
1035
- },
1036
- "llama3_70b_judge": {
1037
- "MERaLiON-AudioLLM-Whisper-SEA-LION": 0.7027027027027027,
1038
- "phi_4_multimodal_instruct": 0.6456456456456456
1039
- }
1040
- },
1041
- "mmau_mini_speech": {
1042
- "string_match": {
1043
- "MERaLiON-AudioLLM-Whisper-SEA-LION": 0.5465465465465466
1044
- },
1045
- "llama3_70b_judge": {
1046
- "MERaLiON-AudioLLM-Whisper-SEA-LION": 0.5915915915915916,
1047
- "phi_4_multimodal_instruct": 0.44744744744744747
1048
- }
1049
- },
1050
  "imda_30s_ar_test": {
1051
  "llama3_70b_judge": {
1052
  "Qwen2-Audio-7B-Instruct": 5.106666666666667,
 
5
  "MERaLiON-AudioLLM-Whisper-SEA-LION": 47.01682396389003,
6
  "Qwen2-Audio-7B-Instruct": 29.187525646286417,
7
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 24.640951990151827,
8
+ "phi_4_multimodal_instruct": 26.815757078375054,
9
  "WavLLM_fairseq": 39.96717275338531,
10
  "SALMONN_7B": 34.222404595814524,
11
  "cascade_whisper_large_v3_llama_3_8b_instruct": 39.32704144439885
 
14
  "cascade_whisper_large_v3_llama_3_8b_instruct": 39.462453836684446
15
  }
16
  },
17
+ "imda_part4_30s_asr_test": {
18
+ "wer": {
19
+ "Qwen-Audio-Chat": 1.1764312018747907,
20
+ "MERaLiON-AudioLLM-Whisper-SEA-LION": 0.0,
21
+ "Qwen2-Audio-7B-Instruct": 0.5685405990059489,
22
+ "whisper_large_v3": 0.8294532718704128,
23
+ "phi_4_multimodal_instruct": 1.3868687388941825
24
+ }
25
+ },
26
  "wavcaps_test": {
27
  "meteor": {
28
  "Qwen-Audio-Chat": 0.2355106805560457,
29
  "MERaLiON-AudioLLM-Whisper-SEA-LION": 0.3175511907248581,
30
  "Qwen2-Audio-7B-Instruct": 0.21342294856199182,
31
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 0.120421856260385,
32
+ "phi_4_multimodal_instruct": 0.24508284335582894,
33
  "WavLLM_fairseq": 0.06399522524688675,
34
  "SALMONN_7B": 0.17175112770658157,
35
  "cascade_whisper_large_v3_llama_3_8b_instruct": 0.1388630786594543
 
39
  "MERaLiON-AudioLLM-Whisper-SEA-LION": 33.97687861271676,
40
  "Qwen2-Audio-7B-Instruct": 33.78034682080925,
41
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 6.3468208092485545,
42
+ "phi_4_multimodal_instruct": 21.884393063583815,
43
  "WavLLM_fairseq": 6.901734104046243,
44
  "SALMONN_7B": 23.76878612716763,
45
  "cascade_whisper_large_v3_llama_3_8b_instruct": 3.445086705202312
 
55
  "Qwen2-Audio-7B-Instruct": 16.466557744958333,
56
  "whisper_large_v3": 14.673689493155793,
57
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 15.209998552437538,
58
+ "phi_4_multimodal_instruct": 22.678131781242936,
59
  "WavLLM_fairseq": 2.368659001743569,
60
  "SALMONN_7B": 5.296039450108202,
61
  "cascade_whisper_large_v3_llama_3_8b_instruct": 14.154700735606419
 
67
  "MERaLiON-AudioLLM-Whisper-SEA-LION": 67.2,
68
  "Qwen2-Audio-7B-Instruct": 53.6,
69
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 71.6,
70
+ "phi_4_multimodal_instruct": 66.2,
71
  "WavLLM_fairseq": 62.199999999999996,
72
  "SALMONN_7B": 46.8,
73
  "cascade_whisper_large_v3_llama_3_8b_instruct": 64.0
 
106
  "MERaLiON-AudioLLM-Whisper-SEA-LION": 57.7927548441449,
107
  "Qwen2-Audio-7B-Instruct": 71.60909856781802,
108
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 51.727042965459134,
109
+ "phi_4_multimodal_instruct": 54.422914911541696,
110
  "WavLLM_fairseq": 44.3133951137321,
111
  "SALMONN_7B": 50.88458298230834,
112
  "cascade_whisper_large_v3_llama_3_8b_instruct": 56.44481887110362
 
131
  "Qwen2-Audio-7B-Instruct": 0.1905689473257041,
132
  "whisper_large_v3": 0.3171008846684522,
133
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 0.32988393799204613,
134
+ "phi_4_multimodal_instruct": 0.3470091713334957,
135
  "WavLLM_fairseq": 0.4463923382842302,
136
  "SALMONN_7B": 0.42346400454508565,
137
  "cascade_whisper_large_v3_llama_3_8b_instruct": 0.31912994075156237
 
144
  "Qwen2-Audio-7B-Instruct": 0.18872219319407232,
145
  "whisper_large_v3": 0.11863959266711877,
146
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 0.11416493424197618,
147
+ "phi_4_multimodal_instruct": 0.15921168191570967,
148
  "WavLLM_fairseq": 0.6447482518259942,
149
  "SALMONN_7B": 0.2577708974886327,
150
  "cascade_whisper_large_v3_llama_3_8b_instruct": 0.11773910240019567
 
169
  "Qwen2-Audio-7B-Instruct": 0.060415760304159495,
170
  "whisper_large_v3": 0.03660128246354058,
171
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 0.05307658841999735,
172
+ "phi_4_multimodal_instruct": 0.03879546787220762,
173
  "WavLLM_fairseq": 0.04798834811886432,
174
  "SALMONN_7B": 0.09671439650443565,
175
  "cascade_whisper_large_v3_llama_3_8b_instruct": 0.03714982881570734
 
182
  "Qwen2-Audio-7B-Instruct": 0.035141660693401744,
183
  "whisper_large_v3": 0.01878749009695552,
184
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 0.032349945297468596,
185
+ "phi_4_multimodal_instruct": 0.0167502923755989,
186
  "WavLLM_fairseq": 0.02103218017882069,
187
  "SALMONN_7B": 0.10270871845172973,
188
  "cascade_whisper_large_v3_llama_3_8b_instruct": 0.018334779492209605
 
195
  "Qwen2-Audio-7B-Instruct": 0.2245352799625317,
196
  "whisper_large_v3": 0.1698509342851144,
197
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 0.1789273082575623,
198
+ "phi_4_multimodal_instruct": 0.14552883606001388,
199
  "WavLLM_fairseq": 0.42541061709652933,
200
  "SALMONN_7B": 0.24872817713464365,
201
  "cascade_whisper_large_v3_llama_3_8b_instruct": 0.17467982364056267
 
207
  "MERaLiON-AudioLLM-Whisper-SEA-LION": 65.6,
208
  "Qwen2-Audio-7B-Instruct": 44.800000000000004,
209
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 72.2,
210
+ "phi_4_multimodal_instruct": 30.8,
211
  "WavLLM_fairseq": 19.2,
212
  "SALMONN_7B": 15.8,
213
  "cascade_whisper_large_v3_llama_3_8b_instruct": 63.0
 
222
  "MERaLiON-AudioLLM-Whisper-SEA-LION": 48.505976095617534,
223
  "Qwen2-Audio-7B-Instruct": 53.98406374501992,
224
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 44.322709163346616,
225
+ "phi_4_multimodal_instruct": 41.03585657370518,
226
  "WavLLM_fairseq": 59.76095617529881,
227
  "SALMONN_7B": 23.804780876494025,
228
  "cascade_whisper_large_v3_llama_3_8b_instruct": 46.713147410358566
 
234
  "MERaLiON-AudioLLM-Whisper-SEA-LION": 59.7093023255814,
235
  "Qwen2-Audio-7B-Instruct": 58.31395348837209,
236
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 73.11046511627907,
237
+ "phi_4_multimodal_instruct": 68.40116279069767,
238
  "WavLLM_fairseq": 58.54651162790698,
239
  "SALMONN_7B": 59.24418604651163,
240
  "cascade_whisper_large_v3_llama_3_8b_instruct": 64.94186046511628
 
262
  "Qwen2-Audio-7B-Instruct": 0.11438872500819404,
263
  "whisper_large_v3": 0.10001863741235596,
264
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 0.10600831614192711,
265
+ "phi_4_multimodal_instruct": 0.08262800367606891,
266
  "WavLLM_fairseq": 0.14533325621300636,
267
  "SALMONN_7B": 0.3062255383962828,
268
  "cascade_whisper_large_v3_llama_3_8b_instruct": 0.09876543209876543
 
274
  "MERaLiON-AudioLLM-Whisper-SEA-LION": 73.66473556344609,
275
  "Qwen2-Audio-7B-Instruct": 64.86264249672958,
276
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 88.61894972902262,
277
+ "phi_4_multimodal_instruct": 77.58549803774996,
278
  "WavLLM_fairseq": 77.64903756307233,
279
  "SALMONN_7B": 66.39506634273968,
280
  "cascade_whisper_large_v3_llama_3_8b_instruct": 83.81984675761541
 
291
  "whisper_large_v3": 0.5377268970583734,
292
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 0.5840399155162387,
293
  "gemini-1.5-flash": 1.1100431601824359,
294
+ "phi_4_multimodal_instruct": 0.8529492791331231,
295
  "WavLLM_fairseq": 1.2204842511249197,
296
  "SALMONN_7B": 1.0189782362484312,
297
  "cascade_whisper_large_v3_llama_3_8b_instruct": 0.507882090054792
 
303
  "MERaLiON-AudioLLM-Whisper-SEA-LION": 46.206896551724135,
304
  "Qwen2-Audio-7B-Instruct": 53.9463601532567,
305
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 56.59003831417625,
306
+ "phi_4_multimodal_instruct": 51.609195402298845,
307
  "WavLLM_fairseq": 51.072796934865906,
308
  "SALMONN_7B": 41.7624521072797,
309
  "cascade_whisper_large_v3_llama_3_8b_instruct": 45.593869731800766
 
315
  "MERaLiON-AudioLLM-Whisper-SEA-LION": 53.2,
316
  "Qwen2-Audio-7B-Instruct": 39.6,
317
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 66.0,
318
+ "phi_4_multimodal_instruct": 43.8,
319
  "WavLLM_fairseq": 46.6,
320
  "SALMONN_7B": 36.6,
321
  "cascade_whisper_large_v3_llama_3_8b_instruct": 53.8
 
330
  "MERaLiON-AudioLLM-Whisper-SEA-LION": 99.75379565038982,
331
  "Qwen2-Audio-7B-Instruct": 99.1177677472302,
332
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 34.94050061551087,
333
+ "phi_4_multimodal_instruct": 94.54247025030776,
334
  "WavLLM_fairseq": 69.61427985227739,
335
  "SALMONN_7B": 88.79770209273697,
336
  "cascade_whisper_large_v3_llama_3_8b_instruct": 42.921624948707425
 
342
  "MERaLiON-AudioLLM-Whisper-SEA-LION": 93.76666666666667,
343
  "Qwen2-Audio-7B-Instruct": 61.56666666666667,
344
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 19.6,
345
+ "phi_4_multimodal_instruct": 36.833333333333336,
346
  "WavLLM_fairseq": 46.766666666666666,
347
  "SALMONN_7B": 42.733333333333334,
348
  "cascade_whisper_large_v3_llama_3_8b_instruct": 25.433333333333337
 
367
  "Qwen2-Audio-7B-Instruct": 0.27856006770658537,
368
  "whisper_large_v3": 0.2143555471246589,
369
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 0.22881615619208825,
370
+ "phi_4_multimodal_instruct": 0.22801359968481416,
371
  "WavLLM_fairseq": 0.39796588405247263,
372
  "SALMONN_7B": 0.34868891450584405,
373
  "cascade_whisper_large_v3_llama_3_8b_instruct": 0.22004640235805695
 
391
  "Qwen2-Audio-7B-Instruct": 0.23542555661330924,
392
  "whisper_large_v3": 0.15887899737116104,
393
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 0.1448629161356777,
394
+ "phi_4_multimodal_instruct": 0.24134627375003423,
395
  "WavLLM_fairseq": 0.6671766188447099,
396
  "SALMONN_7B": 0.3597423676988383,
397
  "cascade_whisper_large_v3_llama_3_8b_instruct": 0.15611126487402763
 
428
  "Qwen2-Audio-7B-Instruct": 6.326113431899141,
429
  "whisper_large_v3": 46.01512198258627,
430
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 46.80524126004861,
431
+ "phi_4_multimodal_instruct": 0.36465303013961253,
432
  "WavLLM_fairseq": 5.933522277713613,
433
  "SALMONN_7B": 26.89649039333571,
434
  "cascade_whisper_large_v3_llama_3_8b_instruct": 46.79924664837527
 
465
  "Qwen2-Audio-7B-Instruct": 0.11723812890302816,
466
  "whisper_large_v3": 0.09459022434812692,
467
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 0.09948381629977261,
468
+ "phi_4_multimodal_instruct": 0.09672866386388193,
469
  "WavLLM_fairseq": 0.15491778414546403,
470
  "SALMONN_7B": 0.10765150204693537,
471
  "cascade_whisper_large_v3_llama_3_8b_instruct": 0.09515429104337297
 
489
  "MERaLiON-AudioLLM-Whisper-SEA-LION": 48.4,
490
  "Qwen2-Audio-7B-Instruct": 33.8,
491
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 45.4,
492
+ "phi_4_multimodal_instruct": 41.2,
493
  "WavLLM_fairseq": 31.6,
494
  "SALMONN_7B": 9.0,
495
  "cascade_whisper_large_v3_llama_3_8b_instruct": 37.400000000000006
 
504
  "MERaLiON-AudioLLM-Whisper-SEA-LION": 77.83333333333333,
505
  "Qwen2-Audio-7B-Instruct": 0.9666666666666667,
506
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 7.633333333333334,
507
+ "phi_4_multimodal_instruct": 0.5333333333333333,
508
  "WavLLM_fairseq": 0.23333333333333336,
509
  "SALMONN_7B": 0.06666666666666667,
510
  "cascade_whisper_large_v3_llama_3_8b_instruct": 9.666666666666666
 
516
  "MERaLiON-AudioLLM-Whisper-SEA-LION": 93.48605577689243,
517
  "Qwen2-Audio-7B-Instruct": 92.80876494023903,
518
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 15.737051792828685,
519
+ "phi_4_multimodal_instruct": 59.46215139442231,
520
  "WavLLM_fairseq": 51.932270916334666,
521
  "SALMONN_7B": 81.31474103585658,
522
  "cascade_whisper_large_v3_llama_3_8b_instruct": 44.22310756972111
 
529
  "Qwen2-Audio-7B-Instruct": 0.2080008649583739,
530
  "whisper_large_v3": 0.17210509244242622,
531
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 0.2192622950819672,
532
+ "phi_4_multimodal_instruct": 0.23849064763758243,
533
  "WavLLM_fairseq": 0.48091685587631094,
534
  "SALMONN_7B": 0.3238620391393664,
535
  "cascade_whisper_large_v3_llama_3_8b_instruct": 0.23561466104443723
 
554
  "Qwen2-Audio-7B-Instruct": 74.7247908410392,
555
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 91.85380889476001,
556
  "gemini-1.5-flash": 89.25583443416997,
557
+ "phi_4_multimodal_instruct": 73.18361955085865,
558
  "WavLLM_fairseq": 66.31439894319684,
559
  "SALMONN_7B": 50.99075297225891,
560
  "cascade_whisper_large_v3_llama_3_8b_instruct": 85.2928225451343
 
566
  "MERaLiON-AudioLLM-Whisper-SEA-LION": 84.31782540512285,
567
  "Qwen2-Audio-7B-Instruct": 66.49242028227914,
568
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 89.33612127548353,
569
+ "phi_4_multimodal_instruct": 72.60846837428123,
570
  "WavLLM_fairseq": 66.5446941975954,
571
  "SALMONN_7B": 56.455828541557764,
572
  "cascade_whisper_large_v3_llama_3_8b_instruct": 86.4610559330894
 
578
  "MERaLiON-AudioLLM-Whisper-SEA-LION": 57.0,
579
  "Qwen2-Audio-7B-Instruct": 40.4,
580
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 58.0,
581
+ "phi_4_multimodal_instruct": 52.199999999999996,
582
  "WavLLM_fairseq": 45.199999999999996,
583
  "SALMONN_7B": 17.2,
584
  "cascade_whisper_large_v3_llama_3_8b_instruct": 49.0
 
594
  "Qwen2-Audio-7B-Instruct": 0.09260359129694522,
595
  "whisper_large_v3": 0.12359684029221357,
596
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 0.20886539565639167,
597
+ "phi_4_multimodal_instruct": 0.07466690423868068,
598
  "WavLLM_fairseq": 0.7054601967888183,
599
  "SALMONN_7B": 0.8259290055631446,
600
  "cascade_whisper_large_v3_llama_3_8b_instruct": 0.12450753301261111
 
606
  "MERaLiON-AudioLLM-Whisper-SEA-LION": 51.4,
607
  "Qwen2-Audio-7B-Instruct": 42.0,
608
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 56.0,
609
+ "phi_4_multimodal_instruct": 43.8,
610
  "WavLLM_fairseq": 45.199999999999996,
611
  "SALMONN_7B": 40.599999999999994,
612
  "cascade_whisper_large_v3_llama_3_8b_instruct": 49.0
 
621
  "MERaLiON-AudioLLM-Whisper-SEA-LION": 46.4,
622
  "Qwen2-Audio-7B-Instruct": 24.8,
623
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 44.0,
624
+ "phi_4_multimodal_instruct": 37.0,
625
  "WavLLM_fairseq": 31.6,
626
  "SALMONN_7B": 7.0,
627
  "cascade_whisper_large_v3_llama_3_8b_instruct": 36.0
 
648
  "MERaLiON-AudioLLM-Whisper-SEA-LION": 38.00454545454545,
649
  "Qwen2-Audio-7B-Instruct": 40.77727272727273,
650
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 3.0954545454545457,
651
+ "phi_4_multimodal_instruct": 26.386363636363637,
652
  "WavLLM_fairseq": 5.5,
653
  "SALMONN_7B": 37.445454545454545,
654
  "cascade_whisper_large_v3_llama_3_8b_instruct": 2.4727272727272727
 
658
  "MERaLiON-AudioLLM-Whisper-SEA-LION": 0.24920047034353812,
659
  "Qwen2-Audio-7B-Instruct": 0.19891712076314283,
660
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 0.05796819723943051,
661
+ "phi_4_multimodal_instruct": 0.1757379026471828,
662
  "WavLLM_fairseq": 0.041732965094428545,
663
  "SALMONN_7B": 0.20994052484339956,
664
  "cascade_whisper_large_v3_llama_3_8b_instruct": 0.07953048457785493
 
673
  "MERaLiON-AudioLLM-Whisper-SEA-LION": 7.816666666666666,
674
  "Qwen2-Audio-7B-Instruct": 2.55,
675
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 26.016666666666666,
676
+ "phi_4_multimodal_instruct": 3.5166666666666666,
677
  "WavLLM_fairseq": 2.6833333333333336,
678
  "SALMONN_7B": 2.5166666666666666,
679
  "cascade_whisper_large_v3_llama_3_8b_instruct": 12.416666666666666
 
698
  "Qwen2-Audio-7B-Instruct": 0.04425838146050298,
699
  "whisper_large_v3": 2.451098639578599,
700
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 2.8327095799289337,
701
+ "phi_4_multimodal_instruct": 0.053138495633157125,
702
  "WavLLM_fairseq": 0.1695522548322915,
703
  "SALMONN_7B": 0.3649023706010388,
704
  "cascade_whisper_large_v3_llama_3_8b_instruct": 2.4245628096245917
 
711
  "Qwen2-Audio-7B-Instruct": 16.325186897428104,
712
  "whisper_large_v3": 1.600581653970121,
713
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 27.620150160643625,
714
+ "phi_4_multimodal_instruct": 15.012558278964478,
715
  "WavLLM_fairseq": 13.841886973016162,
716
  "SALMONN_7B": 14.102682915273142,
717
  "cascade_whisper_large_v3_llama_3_8b_instruct": 10.930203684508578
 
723
  "MERaLiON-AudioLLM-Whisper-SEA-LION": 63.15021876519203,
724
  "Qwen2-Audio-7B-Instruct": 50.919591292758774,
725
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 24.647544968400585,
726
+ "phi_4_multimodal_instruct": 47.86582401555663,
727
  "WavLLM_fairseq": 43.01199466903598,
728
  "SALMONN_7B": 57.75401069518716,
729
  "cascade_whisper_large_v3_llama_3_8b_instruct": 29.47134606841404
 
740
  "whisper_large_v3": 0.12226319428439733,
741
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 0.1400092187139894,
742
  "gemini-1.5-flash": 0.1089344703080587,
743
+ "phi_4_multimodal_instruct": 0.16175001920565416,
744
  "WavLLM_fairseq": 0.41876008296842593,
745
  "SALMONN_7B": 0.21487285856956287,
746
  "cascade_whisper_large_v3_llama_3_8b_instruct": 0.12579703464700007
 
753
  "Qwen2-Audio-7B-Instruct": 0.35076166942732234,
754
  "whisper_large_v3": 0.27026366524560785,
755
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 0.3035544573275043,
756
+ "phi_4_multimodal_instruct": 0.44227061666711925,
757
  "WavLLM_fairseq": 0.7540934640345399,
758
  "SALMONN_7B": 0.6569229098215983,
759
  "cascade_whisper_large_v3_llama_3_8b_instruct": 0.29992939962527493
 
765
  "MERaLiON-AudioLLM-Whisper-SEA-LION": 74.80000000000001,
766
  "Qwen2-Audio-7B-Instruct": 52.599999999999994,
767
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 73.8,
768
+ "phi_4_multimodal_instruct": 25.8,
769
  "WavLLM_fairseq": 21.6,
770
  "SALMONN_7B": 17.2,
771
  "cascade_whisper_large_v3_llama_3_8b_instruct": 70.8
 
803
  "MERaLiON-AudioLLM-Whisper-SEA-LION": 49.77635782747604,
804
  "Qwen2-Audio-7B-Instruct": 45.75079872204473,
805
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 18.466453674121407,
806
+ "phi_4_multimodal_instruct": 38.466453674121404,
807
  "WavLLM_fairseq": 29.840255591054312,
808
  "SALMONN_7B": 50.287539936102235,
809
  "cascade_whisper_large_v3_llama_3_8b_instruct": 17.380191693290733
 
819
  "Qwen2-Audio-7B-Instruct": 0.07197717796796138,
820
  "whisper_large_v3": 0.06844171360300393,
821
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 0.07041669714480775,
822
+ "phi_4_multimodal_instruct": 0.05739643527661961,
823
  "WavLLM_fairseq": 0.10077292565771828,
824
  "SALMONN_7B": 0.0925804013361617,
825
  "cascade_whisper_large_v3_llama_3_8b_instruct": 0.06922195401458074
 
844
  "Qwen2-Audio-7B-Instruct": 0.03245972071872916,
845
  "whisper_large_v3": 0.02107778621423822,
846
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 8.433062902024755,
847
+ "phi_4_multimodal_instruct": 0.19835914151649442,
848
  "WavLLM_fairseq": 0.0033159224040994286,
849
  "SALMONN_7B": 0.00046745670226766583,
850
  "cascade_whisper_large_v3_llama_3_8b_instruct": 1.0368044741318085
 
856
  "MERaLiON-AudioLLM-Whisper-SEA-LION": 46.31578947368421,
857
  "Qwen2-Audio-7B-Instruct": 44.473684210526315,
858
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 18.88157894736842,
859
+ "phi_4_multimodal_instruct": 35.13157894736842,
860
  "WavLLM_fairseq": 26.25,
861
  "SALMONN_7B": 47.30263157894737,
862
  "cascade_whisper_large_v3_llama_3_8b_instruct": 16.710526315789473
 
884
  "Qwen2-Audio-7B-Instruct": 25.765420247070075,
885
  "whisper_large_v3": 0.16408986541757878,
886
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 35.274306071307024,
887
+ "phi_4_multimodal_instruct": 45.295964957544776,
888
  "WavLLM_fairseq": 31.96381187282953,
889
  "SALMONN_7B": 33.88941292215531,
890
  "cascade_whisper_large_v3_llama_3_8b_instruct": 5.987143868370054
891
  }
892
  },
893
+ "mmau_mini": {
894
+ "string_match": {
895
+ "Qwen-Audio-Chat": 38.5,
896
+ "MERaLiON-AudioLLM-Whisper-SEA-LION": 60.5,
897
+ "Qwen2-Audio-7B-Instruct": 44.4,
898
+ "phi_4_multimodal_instruct": 54.50000000000001
899
+ },
900
+ "llama3_70b_judge": {
901
+ "Qwen-Audio-Chat": 53.6,
902
+ "MERaLiON-AudioLLM-Whisper-SEA-LION": 64.60000000000001,
903
+ "Qwen2-Audio-7B-Instruct": 58.9,
904
+ "phi_4_multimodal_instruct": 59.4
905
+ }
906
+ },
907
+ "mmau_mini_music": {
908
+ "string_match": {
909
+ "Qwen-Audio-Chat": 0.4311377245508982,
910
+ "MERaLiON-AudioLLM-Whisper-SEA-LION": 0.6077844311377245,
911
+ "Qwen2-Audio-7B-Instruct": 0.45808383233532934,
912
+ "phi_4_multimodal_instruct": 0.6377245508982036
913
+ },
914
+ "llama3_70b_judge": {
915
+ "Qwen-Audio-Chat": 0.5958083832335329,
916
+ "MERaLiON-AudioLLM-Whisper-SEA-LION": 0.6437125748502994,
917
+ "Qwen2-Audio-7B-Instruct": 0.6017964071856288,
918
+ "phi_4_multimodal_instruct": 0.688622754491018
919
+ }
920
+ },
921
+ "mmau_mini_sound": {
922
+ "string_match": {
923
+ "Qwen-Audio-Chat": 0.43543543543543545,
924
+ "MERaLiON-AudioLLM-Whisper-SEA-LION": 0.6606606606606606,
925
+ "Qwen2-Audio-7B-Instruct": 0.4744744744744745,
926
+ "phi_4_multimodal_instruct": 0.5975975975975976
927
+ },
928
+ "llama3_70b_judge": {
929
+ "Qwen-Audio-Chat": 0.5945945945945946,
930
+ "MERaLiON-AudioLLM-Whisper-SEA-LION": 0.7027027027027027,
931
+ "Qwen2-Audio-7B-Instruct": 0.6306306306306306,
932
+ "phi_4_multimodal_instruct": 0.6456456456456456
933
+ }
934
+ },
935
+ "mmau_mini_speech": {
936
+ "string_match": {
937
+ "Qwen-Audio-Chat": 0.2882882882882883,
938
+ "MERaLiON-AudioLLM-Whisper-SEA-LION": 0.5465465465465466,
939
+ "Qwen2-Audio-7B-Instruct": 0.3993993993993994,
940
+ "phi_4_multimodal_instruct": 0.3993993993993994
941
+ },
942
+ "llama3_70b_judge": {
943
+ "Qwen-Audio-Chat": 0.4174174174174174,
944
+ "MERaLiON-AudioLLM-Whisper-SEA-LION": 0.5915915915915916,
945
+ "Qwen2-Audio-7B-Instruct": 0.5345345345345346,
946
+ "phi_4_multimodal_instruct": 0.44744744744744747
947
+ }
948
+ },
949
  "imda_part5_30s_sqa_test": {
950
  "llama3_70b_judge": {
951
  "Qwen-Audio-Chat": 61.260000000000005,
 
964
  "MERaLiON-AudioLLM-Whisper-SEA-LION": 66.13333333333333,
965
  "Qwen2-Audio-7B-Instruct": 68.38333333333333,
966
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 26.35,
967
+ "phi_4_multimodal_instruct": 51.68333333333334,
968
  "WavLLM_fairseq": 49.06666666666666,
969
  "SALMONN_7B": 59.766666666666666,
970
  "cascade_whisper_large_v3_llama_3_8b_instruct": 36.016666666666666
 
976
  "MERaLiON-AudioLLM-Whisper-SEA-LION": 86.76470588235293,
977
  "Qwen2-Audio-7B-Instruct": 80.04901960784315,
978
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 88.57843137254902,
979
+ "phi_4_multimodal_instruct": 88.33333333333334,
980
  "WavLLM_fairseq": 83.92156862745098,
981
  "SALMONN_7B": 83.48039215686273,
982
  "cascade_whisper_large_v3_llama_3_8b_instruct": 82.99019607843137
 
992
  "Qwen2-Audio-7B-Instruct": 0.08739585179932637,
993
  "whisper_large_v3": 0.03208650948413402,
994
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 0.04396383619925545,
995
+ "phi_4_multimodal_instruct": 0.0381847190214501,
996
  "WavLLM_fairseq": 0.4536784258110264,
997
  "SALMONN_7B": 0.14231519234178336,
998
  "cascade_whisper_large_v3_llama_3_8b_instruct": 0.04754476156709803
 
1004
  "MERaLiON-AudioLLM-Whisper-SEA-LION": 36.36015325670498,
1005
  "Qwen2-Audio-7B-Instruct": 41.60919540229885,
1006
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 47.356321839080465,
1007
+ "phi_4_multimodal_instruct": 43.524904214559385,
1008
  "WavLLM_fairseq": 41.57088122605364,
1009
  "SALMONN_7B": 30.536398467432953,
1010
  "cascade_whisper_large_v3_llama_3_8b_instruct": 36.81992337164751
 
1017
  "Qwen2-Audio-7B-Instruct": 0.06114048472375004,
1018
  "whisper_large_v3": 0.037649480146197796,
1019
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 0.04900464852205386,
1020
+ "phi_4_multimodal_instruct": 0.028494375643163834,
1021
  "WavLLM_fairseq": 0.06621482559171073,
1022
  "SALMONN_7B": 0.0459884319222171,
1023
  "cascade_whisper_large_v3_llama_3_8b_instruct": 0.038146268762641496
 
1031
  "whisper_large_v3": 0.7225930420711975,
1032
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 0.7824973031283711,
1033
  "gemini-1.5-flash": 0.9690871089536138,
1034
+ "phi_4_multimodal_instruct": 0.7126483279395901,
1035
  "WavLLM_fairseq": 1.2913969795037756,
1036
  "SALMONN_7B": 1.2721817691477886,
1037
  "cascade_whisper_large_v3_llama_3_8b_instruct": 0.6848705501618123
 
1055
  "MERaLiON-AudioLLM-Whisper-SEA-LION": 64.80000000000001,
1056
  "Qwen2-Audio-7B-Instruct": 51.6,
1057
  "cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct": 74.0,
1058
+ "phi_4_multimodal_instruct": 49.0,
1059
  "WavLLM_fairseq": 50.8,
1060
  "SALMONN_7B": 44.6,
1061
  "cascade_whisper_large_v3_llama_3_8b_instruct": 57.800000000000004
 
1127
  "cascade_whisper_large_v3_llama_3_8b_instruct": 63.68000000000001
1128
  }
1129
  },
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1130
  "imda_30s_ar_test": {
1131
  "llama3_70b_judge": {
1132
  "Qwen2-Audio-7B-Instruct": 5.106666666666667,