update seallm-audio and whisper
Browse files- model_information.py +4 -0
- results_organized/bleu/st.csv +2 -0
- results_organized/llama3_70b_judge/accent_recognition.csv +2 -0
- results_organized/llama3_70b_judge/audio_captioning.csv +2 -0
- results_organized/llama3_70b_judge/audio_scene_question_answering.csv +2 -0
- results_organized/llama3_70b_judge/emotion_recognition.csv +2 -0
- results_organized/llama3_70b_judge/gender_recognition.csv +2 -0
- results_organized/llama3_70b_judge/music_understanding.csv +2 -0
- results_organized/llama3_70b_judge/sds_singlish.csv +2 -0
- results_organized/llama3_70b_judge/speech_instruction.csv +2 -0
- results_organized/llama3_70b_judge/sqa_english.csv +2 -0
- results_organized/llama3_70b_judge/sqa_singlish.csv +2 -0
- results_organized/wer/asr_english.csv +13 -12
- results_organized/wer/asr_indonesian.csv +2 -1
- results_organized/wer/asr_malay.csv +2 -1
- results_organized/wer/asr_mandarin.csv +2 -1
- results_organized/wer/asr_private.csv +2 -1
- results_organized/wer/asr_singlish.csv +2 -1
- results_organized/wer/asr_tamil.csv +2 -1
- results_organized/wer/asr_thai.csv +2 -1
- results_organized/wer/asr_vietnamese.csv +2 -1
model_information.py
CHANGED
@@ -64,6 +64,10 @@ data['Original Name'].append('cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_
|
|
64 |
data['Proper Display Name'].append('Cascade: Whisper-Large-v2 / SEA-LIONv3')
|
65 |
data['Link'].append('https://github.com/aisingapore/sealion')
|
66 |
|
|
|
|
|
|
|
|
|
67 |
data['Original Name'].append('whisper_large_v3')
|
68 |
data['Proper Display Name'].append('Whisper-large-v3')
|
69 |
data['Link'].append('https://huggingface.co/openai/whisper-large-v3')
|
|
|
64 |
data['Proper Display Name'].append('Cascade: Whisper-Large-v2 / SEA-LIONv3')
|
65 |
data['Link'].append('https://github.com/aisingapore/sealion')
|
66 |
|
67 |
+
data['Original Name'].append('SeaLLMs-Audio-7B')
|
68 |
+
data['Proper Display Name'].append('Fusion: SeaLLMs-Audio-7B')
|
69 |
+
data['Link'].append('https://huggingface.co/SeaLLMs/SeaLLMs-Audio-7B')
|
70 |
+
|
71 |
data['Original Name'].append('whisper_large_v3')
|
72 |
data['Proper Display Name'].append('Whisper-large-v3')
|
73 |
data['Link'].append('https://huggingface.co/openai/whisper-large-v3')
|
results_organized/bleu/st.csv
CHANGED
@@ -9,9 +9,11 @@ WavLLM_fairseq,13.841886973016162,31.96381187282953,0.0033159224040994,5.9335222
|
|
9 |
MERaLiON-AudioLLM-Whisper-SEA-LION,37.058238343330466,43.96331874536172,13.808713343771569,43.37364836260576,19.55610418584389,4.758175879451736
|
10 |
MERaLiON-AudioLLM-v2-2b,30.658188021678257,40.02820084309168,5.601731502002274,37.77329494766737,16.777825775562142,1.9423083468131173
|
11 |
MERaLiON-AudioLLM-v2-9b,36.242124109428445,43.747307981166834,10.885517678613343,47.85937752036512,22.133726547487697,3.4786390367027833
|
|
|
12 |
Qwen2.5-Omni-3B,22.677415597466005,41.3904266408345,0.11385307865833577,44.70177381121325,21.564197151391852,0.2121024080246949
|
13 |
Qwen2.5-Omni-7B,22.381473837803917,40.43638195419091,0.7240804450352291,43.844763499607055,16.686179809018903,0.05656546920236443
|
14 |
SALMONN_7B,14.193483776951359,33.255550227097565,0.0005121531999434492,27.88515689237341,5.175547389931541,0.40577007761551664
|
|
|
15 |
cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct,27.59161630015759,28.71368811388653,7.474730798912167,46.80524126004861,15.209998552437538,2.8327095799289337
|
16 |
cascade_whisper_large_v3_llama_3_8b_instruct,10.753313930099422,6.089840198985321,1.0029597453865848,46.79744652156276,14.156349261775734,2.4177196689141547
|
17 |
phi_4_multimodal_instruct,14.553644350540432,45.48015814069248,0.14817117451495013,0.37716244197757426,22.330318273444895,0.07320611681035753
|
|
|
9 |
MERaLiON-AudioLLM-Whisper-SEA-LION,37.058238343330466,43.96331874536172,13.808713343771569,43.37364836260576,19.55610418584389,4.758175879451736
|
10 |
MERaLiON-AudioLLM-v2-2b,30.658188021678257,40.02820084309168,5.601731502002274,37.77329494766737,16.777825775562142,1.9423083468131173
|
11 |
MERaLiON-AudioLLM-v2-9b,36.242124109428445,43.747307981166834,10.885517678613343,47.85937752036512,22.133726547487697,3.4786390367027833
|
12 |
+
MERaLiON-AudioLLM-v2-9b-asr,37.63315974431301,45.65689578791446,12.607234197708449,49.106002513200565,20.844768810363885,5.3940943548279074
|
13 |
Qwen2.5-Omni-3B,22.677415597466005,41.3904266408345,0.11385307865833577,44.70177381121325,21.564197151391852,0.2121024080246949
|
14 |
Qwen2.5-Omni-7B,22.381473837803917,40.43638195419091,0.7240804450352291,43.844763499607055,16.686179809018903,0.05656546920236443
|
15 |
SALMONN_7B,14.193483776951359,33.255550227097565,0.0005121531999434492,27.88515689237341,5.175547389931541,0.40577007761551664
|
16 |
+
SeaLLMs-Audio-7B,28.164939943364082,36.44346774159554,0.012032903969232935,43.85735723738928,18.319750661860983,0.06174297068745826
|
17 |
cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct,27.59161630015759,28.71368811388653,7.474730798912167,46.80524126004861,15.209998552437538,2.8327095799289337
|
18 |
cascade_whisper_large_v3_llama_3_8b_instruct,10.753313930099422,6.089840198985321,1.0029597453865848,46.79744652156276,14.156349261775734,2.4177196689141547
|
19 |
phi_4_multimodal_instruct,14.553644350540432,45.48015814069248,0.14817117451495013,0.37716244197757426,22.330318273444895,0.07320611681035753
|
results_organized/llama3_70b_judge/accent_recognition.csv
CHANGED
@@ -9,9 +9,11 @@ WavLLM_fairseq,39.96717275338531,2.6833333333333336,0.2333333333333333
|
|
9 |
MERaLiON-AudioLLM-Whisper-SEA-LION,47.066064833812064,6.333333333333334,78.0
|
10 |
MERaLiON-AudioLLM-v2-2b,66.59827656955272,59.73333333333334,53.833333333333336
|
11 |
MERaLiON-AudioLLM-v2-9b,40.78785391875257,30.325000000000006,54.333333333333336
|
|
|
12 |
Qwen2.5-Omni-3B,0.9027492819039803,0.1,0.4333333333333333
|
13 |
Qwen2.5-Omni-7B,1.661879359868691,0.06666666666666667,0.03333333333333333
|
14 |
SALMONN_7B,31.69881001231022,2.833333333333333,0.2
|
|
|
15 |
cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct,28.00574476815757,38.983333333333334,10.8
|
16 |
cascade_whisper_large_v3_llama_3_8b_instruct,40.29544521953221,13.733333333333334,10.166666666666666
|
17 |
phi_4_multimodal_instruct,2.6261797291752154,6.133333333333333,0.5333333333333333
|
|
|
9 |
MERaLiON-AudioLLM-Whisper-SEA-LION,47.066064833812064,6.333333333333334,78.0
|
10 |
MERaLiON-AudioLLM-v2-2b,66.59827656955272,59.73333333333334,53.833333333333336
|
11 |
MERaLiON-AudioLLM-v2-9b,40.78785391875257,30.325000000000006,54.333333333333336
|
12 |
+
MERaLiON-AudioLLM-v2-9b-asr,18.38325810422651,58.25,66.73333333333333
|
13 |
Qwen2.5-Omni-3B,0.9027492819039803,0.1,0.4333333333333333
|
14 |
Qwen2.5-Omni-7B,1.661879359868691,0.06666666666666667,0.03333333333333333
|
15 |
SALMONN_7B,31.69881001231022,2.833333333333333,0.2
|
16 |
+
SeaLLMs-Audio-7B,9.848173984407058,3.85,16.733333333333334
|
17 |
cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct,28.00574476815757,38.983333333333334,10.8
|
18 |
cascade_whisper_large_v3_llama_3_8b_instruct,40.29544521953221,13.733333333333334,10.166666666666666
|
19 |
phi_4_multimodal_instruct,2.6261797291752154,6.133333333333333,0.5333333333333333
|
results_organized/llama3_70b_judge/audio_captioning.csv
CHANGED
@@ -9,9 +9,11 @@ WavLLM_fairseq,5.5,6.901734104046243
|
|
9 |
MERaLiON-AudioLLM-Whisper-SEA-LION,39.38636363636363,34.566473988439306
|
10 |
MERaLiON-AudioLLM-v2-2b,35.07727272727273,31.410404624277458
|
11 |
MERaLiON-AudioLLM-v2-9b,36.04090909090909,35.16763005780347
|
|
|
12 |
Qwen2.5-Omni-3B,43.69545454545454,34.70520231213873
|
13 |
Qwen2.5-Omni-7B,37.7,26.09248554913295
|
14 |
SALMONN_7B,35.24090909090909,22.520231213872833
|
|
|
15 |
cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct,2.4545454545454546,3.8265895953757223
|
16 |
cascade_whisper_large_v3_llama_3_8b_instruct,2.5136363636363637,3.3179190751445087
|
17 |
phi_4_multimodal_instruct,33.595454545454544,28.069364161849713
|
|
|
9 |
MERaLiON-AudioLLM-Whisper-SEA-LION,39.38636363636363,34.566473988439306
|
10 |
MERaLiON-AudioLLM-v2-2b,35.07727272727273,31.410404624277458
|
11 |
MERaLiON-AudioLLM-v2-9b,36.04090909090909,35.16763005780347
|
12 |
+
MERaLiON-AudioLLM-v2-9b-asr,35.372727272727275,33.5606936416185
|
13 |
Qwen2.5-Omni-3B,43.69545454545454,34.70520231213873
|
14 |
Qwen2.5-Omni-7B,37.7,26.09248554913295
|
15 |
SALMONN_7B,35.24090909090909,22.520231213872833
|
16 |
+
SeaLLMs-Audio-7B,51.95909090909091,38.21965317919075
|
17 |
cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct,2.4545454545454546,3.8265895953757223
|
18 |
cascade_whisper_large_v3_llama_3_8b_instruct,2.5136363636363637,3.3179190751445087
|
19 |
phi_4_multimodal_instruct,33.595454545454544,28.069364161849713
|
results_organized/llama3_70b_judge/audio_scene_question_answering.csv
CHANGED
@@ -9,9 +9,11 @@ WavLLM_fairseq,43.01199466903598,29.840255591054312,26.25
|
|
9 |
MERaLiON-AudioLLM-Whisper-SEA-LION,62.67379679144385,48.81789137380192,45.131578947368425
|
10 |
MERaLiON-AudioLLM-v2-2b,50.53962080700049,44.79233226837061,43.0921052631579
|
11 |
MERaLiON-AudioLLM-v2-9b,58.20126397666505,50.35143769968051,44.868421052631575
|
|
|
12 |
Qwen2.5-Omni-3B,52.64948954788527,48.56230031948882,43.15789473684211
|
13 |
Qwen2.5-Omni-7B,46.592124453087024,50.41533546325879,40.0
|
14 |
SALMONN_7B,58.19154107924162,50.35143769968051,46.90789473684211
|
|
|
15 |
cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct,22.673796791443852,17.44408945686901,14.013157894736842
|
16 |
cascade_whisper_large_v3_llama_3_8b_instruct,29.820126397666506,17.06070287539936,18.75
|
17 |
phi_4_multimodal_instruct,48.37141468157511,40.319488817891376,37.96052631578947
|
|
|
9 |
MERaLiON-AudioLLM-Whisper-SEA-LION,62.67379679144385,48.81789137380192,45.131578947368425
|
10 |
MERaLiON-AudioLLM-v2-2b,50.53962080700049,44.79233226837061,43.0921052631579
|
11 |
MERaLiON-AudioLLM-v2-9b,58.20126397666505,50.35143769968051,44.868421052631575
|
12 |
+
MERaLiON-AudioLLM-v2-9b-asr,56.49975692756441,49.00958466453674,43.02631578947369
|
13 |
Qwen2.5-Omni-3B,52.64948954788527,48.56230031948882,43.15789473684211
|
14 |
Qwen2.5-Omni-7B,46.592124453087024,50.41533546325879,40.0
|
15 |
SALMONN_7B,58.19154107924162,50.35143769968051,46.90789473684211
|
16 |
+
SeaLLMs-Audio-7B,53.89402041808459,54.12140575079873,42.56578947368421
|
17 |
cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct,22.673796791443852,17.44408945686901,14.013157894736842
|
18 |
cascade_whisper_large_v3_llama_3_8b_instruct,29.820126397666506,17.06070287539936,18.75
|
19 |
phi_4_multimodal_instruct,48.37141468157511,40.319488817891376,37.96052631578947
|
results_organized/llama3_70b_judge/emotion_recognition.csv
CHANGED
@@ -9,9 +9,11 @@ WavLLM_fairseq,59.76095617529881,51.072796934865906,41.57088122605364
|
|
9 |
MERaLiON-AudioLLM-Whisper-SEA-LION,49.103585657370516,52.452107279693486,44.17624521072797
|
10 |
MERaLiON-AudioLLM-v2-2b,51.39442231075698,58.582375478927204,52.1455938697318
|
11 |
MERaLiON-AudioLLM-v2-9b,62.54980079681275,68.85057471264368,59.808429118773944
|
|
|
12 |
Qwen2.5-Omni-3B,34.36254980079681,30.421455938697317,34.32950191570881
|
13 |
Qwen2.5-Omni-7B,36.55378486055777,27.77777777777778,30.07662835249042
|
14 |
SALMONN_7B,26.195219123505975,42.26053639846744,32.298850574712645
|
|
|
15 |
cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct,41.98207171314741,58.39080459770115,44.272030651341
|
16 |
cascade_whisper_large_v3_llama_3_8b_instruct,46.91235059760957,56.47509578544061,49.42528735632184
|
17 |
phi_4_multimodal_instruct,32.07171314741036,49.11877394636016,40.84291187739464
|
|
|
9 |
MERaLiON-AudioLLM-Whisper-SEA-LION,49.103585657370516,52.452107279693486,44.17624521072797
|
10 |
MERaLiON-AudioLLM-v2-2b,51.39442231075698,58.582375478927204,52.1455938697318
|
11 |
MERaLiON-AudioLLM-v2-9b,62.54980079681275,68.85057471264368,59.808429118773944
|
12 |
+
MERaLiON-AudioLLM-v2-9b-asr,53.685258964143436,59.84674329501916,47.547892720306514
|
13 |
Qwen2.5-Omni-3B,34.36254980079681,30.421455938697317,34.32950191570881
|
14 |
Qwen2.5-Omni-7B,36.55378486055777,27.77777777777778,30.07662835249042
|
15 |
SALMONN_7B,26.195219123505975,42.26053639846744,32.298850574712645
|
16 |
+
SeaLLMs-Audio-7B,51.69322709163347,50.804597701149426,53.524904214559385
|
17 |
cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct,41.98207171314741,58.39080459770115,44.272030651341
|
18 |
cascade_whisper_large_v3_llama_3_8b_instruct,46.91235059760957,56.47509578544061,49.42528735632184
|
19 |
phi_4_multimodal_instruct,32.07171314741036,49.11877394636016,40.84291187739464
|
results_organized/llama3_70b_judge/gender_recognition.csv
CHANGED
@@ -9,9 +9,11 @@ WavLLM_fairseq,69.61427985227739,51.932270916334666
|
|
9 |
MERaLiON-AudioLLM-Whisper-SEA-LION,99.73327862125564,94.6215139442231
|
10 |
MERaLiON-AudioLLM-v2-2b,99.69224456298728,87.92828685258964
|
11 |
MERaLiON-AudioLLM-v2-9b,97.2507180960197,92.96812749003983
|
|
|
12 |
Qwen2.5-Omni-3B,32.78621255642183,62.948207171314735
|
13 |
Qwen2.5-Omni-7B,54.08288879770209,43.366533864541836
|
14 |
SALMONN_7B,88.53098071399262,80.199203187251
|
|
|
15 |
cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct,26.631103816167418,12.211155378486056
|
16 |
cascade_whisper_large_v3_llama_3_8b_instruct,69.69634796881412,44.38247011952191
|
17 |
phi_4_multimodal_instruct,94.58350430857611,46.852589641434264
|
|
|
9 |
MERaLiON-AudioLLM-Whisper-SEA-LION,99.73327862125564,94.6215139442231
|
10 |
MERaLiON-AudioLLM-v2-2b,99.69224456298728,87.92828685258964
|
11 |
MERaLiON-AudioLLM-v2-9b,97.2507180960197,92.96812749003983
|
12 |
+
MERaLiON-AudioLLM-v2-9b-asr,99.26138695116947,95.1792828685259
|
13 |
Qwen2.5-Omni-3B,32.78621255642183,62.948207171314735
|
14 |
Qwen2.5-Omni-7B,54.08288879770209,43.366533864541836
|
15 |
SALMONN_7B,88.53098071399262,80.199203187251
|
16 |
+
SeaLLMs-Audio-7B,82.2732868280673,68.62549800796812
|
17 |
cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct,26.631103816167418,12.211155378486056
|
18 |
cascade_whisper_large_v3_llama_3_8b_instruct,69.69634796881412,44.38247011952191
|
19 |
phi_4_multimodal_instruct,94.58350430857611,46.852589641434264
|
results_organized/llama3_70b_judge/music_understanding.csv
CHANGED
@@ -9,9 +9,11 @@ WavLLM_fairseq,44.3133951137321
|
|
9 |
MERaLiON-AudioLLM-Whisper-SEA-LION,51.34793597304128
|
10 |
MERaLiON-AudioLLM-v2-2b,55.602358887952825
|
11 |
MERaLiON-AudioLLM-v2-9b,63.94271272114573
|
|
|
12 |
Qwen2.5-Omni-3B,59.30918281381634
|
13 |
Qwen2.5-Omni-7B,47.598989048020215
|
14 |
SALMONN_7B,49.70513900589722
|
|
|
15 |
cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct,50.463352990732936
|
16 |
cascade_whisper_large_v3_llama_3_8b_instruct,56.697556866048856
|
17 |
phi_4_multimodal_instruct,55.2653748946925
|
|
|
9 |
MERaLiON-AudioLLM-Whisper-SEA-LION,51.34793597304128
|
10 |
MERaLiON-AudioLLM-v2-2b,55.602358887952825
|
11 |
MERaLiON-AudioLLM-v2-9b,63.94271272114573
|
12 |
+
MERaLiON-AudioLLM-v2-9b-asr,60.65711878685762
|
13 |
Qwen2.5-Omni-3B,59.30918281381634
|
14 |
Qwen2.5-Omni-7B,47.598989048020215
|
15 |
SALMONN_7B,49.70513900589722
|
16 |
+
SeaLLMs-Audio-7B,63.689974726200504
|
17 |
cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct,50.463352990732936
|
18 |
cascade_whisper_large_v3_llama_3_8b_instruct,56.697556866048856
|
19 |
phi_4_multimodal_instruct,55.2653748946925
|
results_organized/llama3_70b_judge/sds_singlish.csv
CHANGED
@@ -9,9 +9,11 @@ WavLLM_fairseq,31.6,31.6,45.2,49.400000000000006
|
|
9 |
MERaLiON-AudioLLM-Whisper-SEA-LION,47.800000000000004,46.4,54.6,65.6
|
10 |
MERaLiON-AudioLLM-v2-2b,42.2,40.199999999999996,51.8,60.0
|
11 |
MERaLiON-AudioLLM-v2-9b,49.8,46.6,55.4,60.599999999999994
|
|
|
12 |
Qwen2.5-Omni-3B,42.800000000000004,33.199999999999996,52.199999999999996,58.8
|
13 |
Qwen2.5-Omni-7B,39.8,31.6,42.800000000000004,58.4
|
14 |
SALMONN_7B,9.0,7.4,16.0,25.2
|
|
|
15 |
cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct,48.4,45.599999999999994,53.4,56.6
|
16 |
cascade_whisper_large_v3_llama_3_8b_instruct,38.0,38.199999999999996,46.2,61.0
|
17 |
phi_4_multimodal_instruct,43.6,42.800000000000004,55.599999999999994,61.0
|
|
|
9 |
MERaLiON-AudioLLM-Whisper-SEA-LION,47.800000000000004,46.4,54.6,65.6
|
10 |
MERaLiON-AudioLLM-v2-2b,42.2,40.199999999999996,51.8,60.0
|
11 |
MERaLiON-AudioLLM-v2-9b,49.8,46.6,55.4,60.599999999999994
|
12 |
+
MERaLiON-AudioLLM-v2-9b-asr,53.0,49.6,58.2,62.400000000000006
|
13 |
Qwen2.5-Omni-3B,42.800000000000004,33.199999999999996,52.199999999999996,58.8
|
14 |
Qwen2.5-Omni-7B,39.8,31.6,42.800000000000004,58.4
|
15 |
SALMONN_7B,9.0,7.4,16.0,25.2
|
16 |
+
SeaLLMs-Audio-7B,42.0,35.2,49.6,55.0
|
17 |
cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct,48.4,45.599999999999994,53.4,56.6
|
18 |
cascade_whisper_large_v3_llama_3_8b_instruct,38.0,38.199999999999996,46.2,61.0
|
19 |
phi_4_multimodal_instruct,43.6,42.800000000000004,55.599999999999994,61.0
|
results_organized/llama3_70b_judge/speech_instruction.csv
CHANGED
@@ -9,9 +9,11 @@ WavLLM_fairseq,19.2,21.6
|
|
9 |
MERaLiON-AudioLLM-Whisper-SEA-LION,66.39999999999999,75.19999999999999
|
10 |
MERaLiON-AudioLLM-v2-2b,12.6,25.6
|
11 |
MERaLiON-AudioLLM-v2-9b,66.2,74.2
|
|
|
12 |
Qwen2.5-Omni-3B,66.0,64.0
|
13 |
Qwen2.5-Omni-7B,57.400000000000006,59.2
|
14 |
SALMONN_7B,15.4,10.4
|
|
|
15 |
cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct,78.8,67.0
|
16 |
cascade_whisper_large_v3_llama_3_8b_instruct,62.800000000000004,69.4
|
17 |
phi_4_multimodal_instruct,39.0,33.4
|
|
|
9 |
MERaLiON-AudioLLM-Whisper-SEA-LION,66.39999999999999,75.19999999999999
|
10 |
MERaLiON-AudioLLM-v2-2b,12.6,25.6
|
11 |
MERaLiON-AudioLLM-v2-9b,66.2,74.2
|
12 |
+
MERaLiON-AudioLLM-v2-9b-asr,10.0,16.8
|
13 |
Qwen2.5-Omni-3B,66.0,64.0
|
14 |
Qwen2.5-Omni-7B,57.400000000000006,59.2
|
15 |
SALMONN_7B,15.4,10.4
|
16 |
+
SeaLLMs-Audio-7B,66.2,67.6
|
17 |
cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct,78.8,67.0
|
18 |
cascade_whisper_large_v3_llama_3_8b_instruct,62.800000000000004,69.4
|
19 |
phi_4_multimodal_instruct,39.0,33.4
|
results_organized/llama3_70b_judge/sqa_english.csv
CHANGED
@@ -9,9 +9,11 @@ WavLLM_fairseq,83.92156862745098,58.54651162790698,77.64903756307233,66.31439894
|
|
9 |
MERaLiON-AudioLLM-Whisper-SEA-LION,86.7156862745098,59.59302325581396,74.20669033825453,57.11140466754734,51.54208050182959,53.1
|
10 |
MERaLiON-AudioLLM-v2-2b,83.18627450980392,69.47674418604652,81.4614090824145,66.00616468516073,61.16048092002091,50.99999999999999
|
11 |
MERaLiON-AudioLLM-v2-9b,89.55882352941177,75.02906976744187,89.20949355260699,84.58828709819463,83.32462101411396,56.699999999999996
|
|
|
12 |
Qwen2.5-Omni-3B,73.87254901960785,61.07558139534884,59.8504952345356,81.41787758696609,69.99477260846837,60.699999999999996
|
13 |
Qwen2.5-Omni-7B,77.30392156862746,61.71511627906977,62.86675387777986,81.72611184500221,70.77888133821223,56.10000000000001
|
14 |
SALMONN_7B,80.88235294117646,59.38953488372093,65.64754251541768,50.81461911052399,56.56037637219028,50.6
|
|
|
15 |
cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct,51.51960784313726,70.93023255813954,57.16314707531303,89.52003522677234,85.15420805018296,52.6
|
16 |
cascade_whisper_large_v3_llama_3_8b_instruct,86.96078431372548,69.68023255813954,87.43412446271725,84.98458828709819,86.1996863565081,55.900000000000006
|
17 |
phi_4_multimodal_instruct,83.72549019607844,74.18604651162791,83.19566436180153,75.6494936151475,77.5222164140094,58.8
|
|
|
9 |
MERaLiON-AudioLLM-Whisper-SEA-LION,86.7156862745098,59.59302325581396,74.20669033825453,57.11140466754734,51.54208050182959,53.1
|
10 |
MERaLiON-AudioLLM-v2-2b,83.18627450980392,69.47674418604652,81.4614090824145,66.00616468516073,61.16048092002091,50.99999999999999
|
11 |
MERaLiON-AudioLLM-v2-9b,89.55882352941177,75.02906976744187,89.20949355260699,84.58828709819463,83.32462101411396,56.699999999999996
|
12 |
+
MERaLiON-AudioLLM-v2-9b-asr,84.85294117647057,71.74418604651163,84.71687535040179,72.6552179656539,74.38578149503398,55.5
|
13 |
Qwen2.5-Omni-3B,73.87254901960785,61.07558139534884,59.8504952345356,81.41787758696609,69.99477260846837,60.699999999999996
|
14 |
Qwen2.5-Omni-7B,77.30392156862746,61.71511627906977,62.86675387777986,81.72611184500221,70.77888133821223,56.10000000000001
|
15 |
SALMONN_7B,80.88235294117646,59.38953488372093,65.64754251541768,50.81461911052399,56.56037637219028,50.6
|
16 |
+
SeaLLMs-Audio-7B,82.84313725490196,57.558139534883715,67.09400112128574,81.72611184500221,75.90172503920543,60.4
|
17 |
cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct,51.51960784313726,70.93023255813954,57.16314707531303,89.52003522677234,85.15420805018296,52.6
|
18 |
cascade_whisper_large_v3_llama_3_8b_instruct,86.96078431372548,69.68023255813954,87.43412446271725,84.98458828709819,86.1996863565081,55.900000000000006
|
19 |
phi_4_multimodal_instruct,83.72549019607844,74.18604651162791,83.19566436180153,75.6494936151475,77.5222164140094,58.8
|
results_organized/llama3_70b_judge/sqa_singlish.csv
CHANGED
@@ -9,9 +9,11 @@ WavLLM_fairseq,45.2,46.6,50.8,62.2
|
|
9 |
MERaLiON-AudioLLM-Whisper-SEA-LION,55.199999999999996,50.0,63.0,67.4
|
10 |
MERaLiON-AudioLLM-v2-2b,52.599999999999994,54.6,61.4,70.19999999999999
|
11 |
MERaLiON-AudioLLM-v2-9b,59.400000000000006,63.0,72.0,71.8
|
|
|
12 |
Qwen2.5-Omni-3B,52.400000000000006,54.400000000000006,66.0,69.2
|
13 |
Qwen2.5-Omni-7B,54.2,52.0,62.800000000000004,64.6
|
14 |
SALMONN_7B,42.0,35.4,45.8,49.6
|
|
|
15 |
cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct,51.4,46.4,54.6,62.599999999999994
|
16 |
cascade_whisper_large_v3_llama_3_8b_instruct,51.6,55.599999999999994,62.0,68.2
|
17 |
phi_4_multimodal_instruct,55.0,56.4,64.6,71.8
|
|
|
9 |
MERaLiON-AudioLLM-Whisper-SEA-LION,55.199999999999996,50.0,63.0,67.4
|
10 |
MERaLiON-AudioLLM-v2-2b,52.599999999999994,54.6,61.4,70.19999999999999
|
11 |
MERaLiON-AudioLLM-v2-9b,59.400000000000006,63.0,72.0,71.8
|
12 |
+
MERaLiON-AudioLLM-v2-9b-asr,55.2,59.2,64.4,68.60000000000001
|
13 |
Qwen2.5-Omni-3B,52.400000000000006,54.400000000000006,66.0,69.2
|
14 |
Qwen2.5-Omni-7B,54.2,52.0,62.800000000000004,64.6
|
15 |
SALMONN_7B,42.0,35.4,45.8,49.6
|
16 |
+
SeaLLMs-Audio-7B,44.2,46.2,54.800000000000004,60.199999999999996
|
17 |
cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct,51.4,46.4,54.6,62.599999999999994
|
18 |
cascade_whisper_large_v3_llama_3_8b_instruct,51.6,55.599999999999994,62.0,68.2
|
19 |
phi_4_multimodal_instruct,55.0,56.4,64.6,71.8
|
results_organized/wer/asr_english.csv
CHANGED
@@ -4,15 +4,16 @@ Qwen2-Audio-7B-Instruct,0.0351416606934017,0.0604157603041594,0.114388725008194,
|
|
4 |
old_models,,,,,,,,,
|
5 |
gemini-1.5-flash,,,,,,,,,
|
6 |
WavLLM_fairseq,0.0210321801788206,0.0479883481188643,0.1453332562130063,0.3792176325635977,0.154917784145464,0.6447482518259942,0.6671766188447099,0.0662148255917107,0.4536784258110264
|
7 |
-
MERaLiON-AudioLLM-Whisper-SEA-LION,0.023937073225940318,0.0422569845082944,0.
|
8 |
-
MERaLiON-AudioLLM-v2-2b,0.027124910401026145,0.050958064577146425,0.
|
9 |
-
MERaLiON-AudioLLM-v2-9b,0.02497453502848304,0.046607524542720415,0.
|
10 |
-
MERaLiON-AudioLLM-v2-9b-asr,0.020956728411363035,0.04040327614579984,0.
|
11 |
-
Qwen2.5-Omni-3B,0.021107631946278342,0.04492405470331209,0.
|
12 |
-
Qwen2.5-Omni-7B,0.04404496925340476,0.06877636332683905,0.
|
13 |
-
SALMONN_7B,0.09638963292715132,0.11776722719276675,0.
|
14 |
-
|
15 |
-
|
16 |
-
|
17 |
-
|
18 |
-
|
|
|
|
4 |
old_models,,,,,,,,,
|
5 |
gemini-1.5-flash,,,,,,,,,
|
6 |
WavLLM_fairseq,0.0210321801788206,0.0479883481188643,0.1453332562130063,0.3792176325635977,0.154917784145464,0.6447482518259942,0.6671766188447099,0.0662148255917107,0.4536784258110264
|
7 |
+
MERaLiON-AudioLLM-Whisper-SEA-LION,0.023937073225940318,0.0422569845082944,0.07734234900148476,0.21620323529945748,0.14477210452030514,0.13838923413858656,0.16553574886426656,0.08154430289911642,0.10512320510547775
|
8 |
+
MERaLiON-AudioLLM-v2-2b,0.027124910401026145,0.050958064577146425,0.09276196964925024,0.20627055897299626,0.09237908290276242,0.21886082422652334,0.23935918375209228,0.03456229374401192,0.13837971990781775
|
9 |
+
MERaLiON-AudioLLM-v2-9b,0.02497453502848304,0.046607524542720415,0.09001741857938951,0.20476530792451958,0.09023061553464748,0.1084090226901313,0.15062142184399924,0.03513005216280473,0.043573834426520124
|
10 |
+
MERaLiON-AudioLLM-v2-9b-asr,0.020956728411363035,0.04040327614579984,0.07591543954595999,0.1957668115250735,0.08768103407213536,0.09210848128425476,0.1277414998676963,0.0313686526383024,0.03495834071973054
|
11 |
+
Qwen2.5-Omni-3B,0.021107631946278342,0.04492405470331209,0.08823056799994858,0.2615060102759792,0.11446542772550759,0.14654089448699847,0.19688006593894564,0.04804655619034101,0.07147668853040241
|
12 |
+
Qwen2.5-Omni-7B,0.04404496925340476,0.06877636332683905,0.07962411862631041,0.3124105638254503,0.13967544855837088,0.18939756089426465,0.24105023789319796,0.049146588126752065,0.08381492643148378
|
13 |
+
SALMONN_7B,0.09638963292715132,0.11776722719276675,0.3197948335593678,0.24158949229136512,0.11024871580815716,0.27733154717568453,0.37956460424973665,0.039352755402576205,0.14139336996986349
|
14 |
+
SeaLLMs-Audio-7B,0.05106198362696646,0.0971116197249702,0.15751923435380927,0.37540994158899715,0.12728854393076772,0.3793954486331447,0.45555254788532057,0.047123948759802706,0.08994859067541217
|
15 |
+
cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct,0.03299128532085864,0.05381428868670437,0.10530206130568642,0.20285898669536326,0.09994259054523941,0.14091838890062366,0.17187922953626794,0.04939498243497392,0.08636766530756958
|
16 |
+
cascade_whisper_large_v3_llama_3_8b_instruct,0.018032972422378994,0.035504189759207064,0.09794897834568489,0.14542012514049835,0.09501640807342393,0.10872308256717546,0.1459710229559586,0.038146268762641496,0.04935295160432548
|
17 |
+
hy_whisper_local_cs,0.029086656354925113,0.05591389713810127,0.10600266099330895,0.17879147486544342,0.10212866235970408,0.14925070316060968,0.17014458107377883,0.04666264504453355,0.06973940790639957
|
18 |
+
phi_4_multimodal_instruct,0.016844607084920964,0.03851173700039722,0.07948914070484185,0.2147161396912585,0.0988294989332872,0.1306461295594268,0.22572024408764688,0.028636315247862035,0.05062932104236838
|
19 |
+
whisper_large_v3,0.018938393631870828,0.03631755159173019,0.09819322410834228,0.14557692212471468,0.09593897049053747,0.10783648875498883,0.1409171231397644,0.0382882083673397,0.04555929799680908
|
results_organized/wer/asr_indonesian.csv
CHANGED
@@ -6,8 +6,9 @@ MERaLiON-AudioLLM-v2-9b-asr,0.07921611923820039,0.16282383194620612
|
|
6 |
Qwen2.5-Omni-3B,0.13579906155120067,0.2746338157871875
|
7 |
Qwen2.5-Omni-7B,0.10994571717729322,0.22737303007662502
|
8 |
SALMONN_7B,1.1888858220627472,2.1181172136986777
|
|
|
9 |
cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct,0.09977918851780293,0.2191718937327333
|
10 |
cascade_whisper_large_v3_llama_3_8b_instruct,0.07815806421933941,0.1926224523482703
|
11 |
hy_whisper_local_cs,0.10267733922163952,0.21382030476256667
|
12 |
phi_4_multimodal_instruct,1.327169012788665,5.803850364012302
|
13 |
-
whisper_large_v3,0.
|
|
|
6 |
Qwen2.5-Omni-3B,0.13579906155120067,0.2746338157871875
|
7 |
Qwen2.5-Omni-7B,0.10994571717729322,0.22737303007662502
|
8 |
SALMONN_7B,1.1888858220627472,2.1181172136986777
|
9 |
+
SeaLLMs-Audio-7B,0.12434446591222743,0.31610863030597885
|
10 |
cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct,0.09977918851780293,0.2191718937327333
|
11 |
cascade_whisper_large_v3_llama_3_8b_instruct,0.07815806421933941,0.1926224523482703
|
12 |
hy_whisper_local_cs,0.10267733922163952,0.21382030476256667
|
13 |
phi_4_multimodal_instruct,1.327169012788665,5.803850364012302
|
14 |
+
whisper_large_v3,0.07815806421933941,0.19604538425451323
|
results_organized/wer/asr_malay.csv
CHANGED
@@ -6,8 +6,9 @@ MERaLiON-AudioLLM-v2-9b-asr,0.19463823439076827
|
|
6 |
Qwen2.5-Omni-3B,2.943749725768944
|
7 |
Qwen2.5-Omni-7B,1.4606642973103419
|
8 |
SALMONN_7B,1.0858672282918695
|
|
|
9 |
cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct,0.3143784827344127
|
10 |
cascade_whisper_large_v3_llama_3_8b_instruct,0.3119213724715897
|
11 |
hy_whisper_local_cs,0.2421569917950068
|
12 |
phi_4_multimodal_instruct,3.762932736606555
|
13 |
-
whisper_large_v3,0.
|
|
|
6 |
Qwen2.5-Omni-3B,2.943749725768944
|
7 |
Qwen2.5-Omni-7B,1.4606642973103419
|
8 |
SALMONN_7B,1.0858672282918695
|
9 |
+
SeaLLMs-Audio-7B,0.7655653547452942
|
10 |
cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct,0.3143784827344127
|
11 |
cascade_whisper_large_v3_llama_3_8b_instruct,0.3119213724715897
|
12 |
hy_whisper_local_cs,0.2421569917950068
|
13 |
phi_4_multimodal_instruct,3.762932736606555
|
14 |
+
whisper_large_v3,0.21969198367776754
|
results_organized/wer/asr_mandarin.csv
CHANGED
@@ -11,8 +11,9 @@ MERaLiON-AudioLLM-v2-9b-asr,0.043317297222387204,0.1183419954537208,0.1494223635
|
|
11 |
Qwen2.5-Omni-3B,0.02807309298964582,0.11308069111981474,0.25013248542660305
|
12 |
Qwen2.5-Omni-7B,0.02438082793846885,0.07647567313746625,0.20640169581346052
|
13 |
SALMONN_7B,0.9314703727900854,1.0013340021130595,0.8858293587705353
|
|
|
14 |
cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct,0.20889509215814378,0.31938144990021666,0.3469210386857446
|
15 |
cascade_whisper_large_v3_llama_3_8b_instruct,0.12450753301261111,0.1962263748225777,0.2698675145733969
|
16 |
hy_whisper_local_cs,0.15675793391538476,0.287290695068461,0.27520932697403283
|
17 |
phi_4_multimodal_instruct,0.12232978955079092,0.154221316286565,0.44008479067302597
|
18 |
-
whisper_large_v3,0.
|
|
|
11 |
Qwen2.5-Omni-3B,0.02807309298964582,0.11308069111981474,0.25013248542660305
|
12 |
Qwen2.5-Omni-7B,0.02438082793846885,0.07647567313746625,0.20640169581346052
|
13 |
SALMONN_7B,0.9314703727900854,1.0013340021130595,0.8858293587705353
|
14 |
+
SeaLLMs-Audio-7B,0.17752568747401556,0.08992241443710447,0.661897191308956
|
15 |
cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct,0.20889509215814378,0.31938144990021666,0.3469210386857446
|
16 |
cascade_whisper_large_v3_llama_3_8b_instruct,0.12450753301261111,0.1962263748225777,0.2698675145733969
|
17 |
hy_whisper_local_cs,0.15675793391538476,0.287290695068461,0.27520932697403283
|
18 |
phi_4_multimodal_instruct,0.12232978955079092,0.154221316286565,0.44008479067302597
|
19 |
+
whisper_large_v3,0.12454712834828058,0.19757104895254154,0.19082140964493904
|
results_organized/wer/asr_private.csv
CHANGED
@@ -6,8 +6,9 @@ MERaLiON-AudioLLM-v2-9b-asr,0.12709601623411299,0.14009745910198398,0.1661248930
|
|
6 |
Qwen2.5-Omni-3B,0.17398269785325216,0.2111033762617473,0.19863130881094954,0.1476276141218799,0.1515844495262986,0.10048254328697133,0.09075439591605218,0.1622877775217024,0.24454535625473023
|
7 |
Qwen2.5-Omni-7B,0.18322118978959734,0.41385311521058127,0.22035928143712574,0.14144366988981336,0.23546226723293043,0.11022802535717664,0.176214785403668,0.17361911346700468,0.35052438101416367
|
8 |
SALMONN_7B,0.1492577165438428,0.2398190045248869,0.5414884516680923,0.19901056892286936,0.3636883371447239,0.20430031223389156,0.191869918699187,0.2207497887378044,0.3495513028435506
|
|
|
9 |
cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct,0.15171419416853574,0.19735468151757746,0.17040205303678357,0.1541488644029683,0.15754655341391702,0.09007474690131517,0.12278313480809226,0.12475992932319274,0.12552708400908205
|
10 |
cascade_whisper_large_v3_llama_3_8b_instruct,0.13815016554523124,0.15344926428434932,0.16184773310521813,0.11434675061839443,0.15125775890231952,0.06537988456807645,0.08943089430894309,0.10816624414227549,0.08387933830684398
|
11 |
hy_whisper_local_cs,0.14674783723165652,0.18308388444135051,0.17570573139435414,0.12885091072633237,0.1256125449199608,0.07257072570725707,0.16948383437322745,0.1284858262272413,0.14315061087685155
|
12 |
phi_4_multimodal_instruct,0.19080422941364947,0.5388096066829099,0.26073567151411464,0.1217674836968743,0.19813786344331918,0.2778645094143249,0.07521270561542824,0.16939386955519706,0.23232781922369986
|
13 |
-
whisper_large_v3,0.
|
|
|
6 |
Qwen2.5-Omni-3B,0.17398269785325216,0.2111033762617473,0.19863130881094954,0.1476276141218799,0.1515844495262986,0.10048254328697133,0.09075439591605218,0.1622877775217024,0.24454535625473023
|
7 |
Qwen2.5-Omni-7B,0.18322118978959734,0.41385311521058127,0.22035928143712574,0.14144366988981336,0.23546226723293043,0.11022802535717664,0.176214785403668,0.17361911346700468,0.35052438101416367
|
8 |
SALMONN_7B,0.1492577165438428,0.2398190045248869,0.5414884516680923,0.19901056892286936,0.3636883371447239,0.20430031223389156,0.191869918699187,0.2207497887378044,0.3495513028435506
|
9 |
+
SeaLLMs-Audio-7B,0.27347004165331623,0.7187608771319178,1.1654405474764755,0.18754216325612771,0.3995426331264293,0.19386886176554075,0.17613915674040462,0.4502573557655374,0.9038382527840848
|
10 |
cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct,0.15171419416853574,0.19735468151757746,0.17040205303678357,0.1541488644029683,0.15754655341391702,0.09007474690131517,0.12278313480809226,0.12475992932319274,0.12552708400908205
|
11 |
cascade_whisper_large_v3_llama_3_8b_instruct,0.13815016554523124,0.15344926428434932,0.16184773310521813,0.11434675061839443,0.15125775890231952,0.06537988456807645,0.08943089430894309,0.10816624414227549,0.08387933830684398
|
12 |
hy_whisper_local_cs,0.14674783723165652,0.18308388444135051,0.17570573139435414,0.12885091072633237,0.1256125449199608,0.07257072570725707,0.16948383437322745,0.1284858262272413,0.14315061087685155
|
13 |
phi_4_multimodal_instruct,0.19080422941364947,0.5388096066829099,0.26073567151411464,0.1217674836968743,0.19813786344331918,0.2778645094143249,0.07521270561542824,0.16939386955519706,0.23232781922369986
|
14 |
+
whisper_large_v3,0.14162127523229734,0.18795683954054995,0.16544054747647563,0.1163705869125253,0.11287161058477621,0.06708297852209291,0.06976744186046512,0.10686025966044403,0.08169531841280138
|
results_organized/wer/asr_singlish.csv
CHANGED
@@ -11,8 +11,9 @@ MERaLiON-AudioLLM-v2-9b-asr,0.04362031550971643,0.054094635175716256,0.196228310
|
|
11 |
Qwen2.5-Omni-3B,0.05298320044863824,0.0947975002029056,0.47520840687539034,1.2504495215581737,0.27988793392771155,0.18302944168994978
|
12 |
Qwen2.5-Omni-7B,0.05291005291005291,0.09410762113464816,0.5354359573139272,1.3034993524374756,0.37375786140578715,0.27471373891697215
|
13 |
SALMONN_7B,0.09275107892619414,0.45783621459297136,0.681280039101746,0.7865181254636674,0.37533379054734356,0.25522053004731987
|
|
|
14 |
cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct,0.07053860970911661,0.3298433568703839,0.2810437993863198,0.4594298934979693,0.21829536997854984,0.17514817745764627
|
15 |
cascade_whisper_large_v3_llama_3_8b_instruct,0.06922195401458074,0.31912994075156237,0.2770250088250468,0.4581096203900464,0.21391778902978215,0.1722411537654032
|
16 |
hy_whisper_local_cs,0.06692999780557385,0.2735167600032465,0.25580416542210876,0.3612895924757007,0.186411988735025,0.14417222500363377
|
17 |
phi_4_multimodal_instruct,0.057615877892375586,0.3451018586153721,0.4381839411301491,1.4697028756805695,0.23859275364433613,0.1439784234241509
|
18 |
-
whisper_large_v3,0.
|
|
|
11 |
Qwen2.5-Omni-3B,0.05298320044863824,0.0947975002029056,0.47520840687539034,1.2504495215581737,0.27988793392771155,0.18302944168994978
|
12 |
Qwen2.5-Omni-7B,0.05291005291005291,0.09410762113464816,0.5354359573139272,1.3034993524374756,0.37375786140578715,0.27471373891697215
|
13 |
SALMONN_7B,0.09275107892619414,0.45783621459297136,0.681280039101746,0.7865181254636674,0.37533379054734356,0.25522053004731987
|
14 |
+
SeaLLMs-Audio-7B,0.12898349303879258,0.2900738576414252,1.195359383061341,1.8651561065774749,0.6310282937648656,0.6653369724963258
|
15 |
cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct,0.07053860970911661,0.3298433568703839,0.2810437993863198,0.4594298934979693,0.21829536997854984,0.17514817745764627
|
16 |
cascade_whisper_large_v3_llama_3_8b_instruct,0.06922195401458074,0.31912994075156237,0.2770250088250468,0.4581096203900464,0.21391778902978215,0.1722411537654032
|
17 |
hy_whisper_local_cs,0.06692999780557385,0.2735167600032465,0.25580416542210876,0.3612895924757007,0.186411988735025,0.14417222500363377
|
18 |
phi_4_multimodal_instruct,0.057615877892375586,0.3451018586153721,0.4381839411301491,1.4697028756805695,0.23859275364433613,0.1439784234241509
|
19 |
+
whisper_large_v3,0.06919757150171897,0.319251684116549,0.26719526434409535,0.4567139031045279,0.2108972581751324,0.17041618889193946
|
results_organized/wer/asr_tamil.csv
CHANGED
@@ -6,8 +6,9 @@ MERaLiON-AudioLLM-v2-9b-asr,0.12871226564172622,0.1383345045678145,0.54678940715
|
|
6 |
Qwen2.5-Omni-3B,0.8307319012713203,1.653935347856641,1.4607630222683219
|
7 |
Qwen2.5-Omni-7B,0.8465494917777076,0.8666549543218552,1.3615441962983372
|
8 |
SALMONN_7B,1.4272941368377052,1.507519325368939,0.985267900554277
|
|
|
9 |
cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct,0.23805397249380653,0.2724525650035137,0.9665002755178114
|
10 |
cascade_whisper_large_v3_llama_3_8b_instruct,0.24404355317218387,0.28397751229796203,0.8976532365239376
|
11 |
hy_whisper_local_cs,0.31793713743921215,0.33113141250878425,0.8339924151567211
|
12 |
phi_4_multimodal_instruct,1.1784589191228196,1.7016514406184118,2.7500567242552916
|
13 |
-
whisper_large_v3,0.
|
|
|
6 |
Qwen2.5-Omni-3B,0.8307319012713203,1.653935347856641,1.4607630222683219
|
7 |
Qwen2.5-Omni-7B,0.8465494917777076,0.8666549543218552,1.3615441962983372
|
8 |
SALMONN_7B,1.4272941368377052,1.507519325368939,0.985267900554277
|
9 |
+
SeaLLMs-Audio-7B,1.2968793010286785,2.061876317638791,3.6174516223137014
|
10 |
cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct,0.23805397249380653,0.2724525650035137,0.9665002755178114
|
11 |
cascade_whisper_large_v3_llama_3_8b_instruct,0.24404355317218387,0.28397751229796203,0.8976532365239376
|
12 |
hy_whisper_local_cs,0.31793713743921215,0.33113141250878425,0.8339924151567211
|
13 |
phi_4_multimodal_instruct,1.1784589191228196,1.7016514406184118,2.7500567242552916
|
14 |
+
whisper_large_v3,0.24484386310112452,0.2314476458186929,0.8481572720495284
|
results_organized/wer/asr_thai.csv
CHANGED
@@ -6,8 +6,9 @@ MERaLiON-AudioLLM-v2-9b-asr,0.18238237758889023,0.010670019759295851
|
|
6 |
Qwen2.5-Omni-3B,0.3000742248294026,0.026225974492545358
|
7 |
Qwen2.5-Omni-7B,0.23150963725607565,0.021483743488413868
|
8 |
SALMONN_7B,1.2470441757452413,1.1351535836177475
|
|
|
9 |
cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct,0.276058900993655,0.06815160768816239
|
10 |
cascade_whisper_large_v3_llama_3_8b_instruct,0.20872022028013887,0.031794503323154304
|
11 |
hy_whisper_local_cs,0.26486292350053875,0.076019400035926
|
12 |
phi_4_multimodal_instruct,1.7344522925894887,1.2856834920064666
|
13 |
-
whisper_large_v3,0.
|
|
|
6 |
Qwen2.5-Omni-3B,0.3000742248294026,0.026225974492545358
|
7 |
Qwen2.5-Omni-7B,0.23150963725607565,0.021483743488413868
|
8 |
SALMONN_7B,1.2470441757452413,1.1351535836177475
|
9 |
+
SeaLLMs-Audio-7B,0.2088686699389441,0.025435602658523442
|
10 |
cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct,0.276058900993655,0.06815160768816239
|
11 |
cascade_whisper_large_v3_llama_3_8b_instruct,0.20872022028013887,0.031794503323154304
|
12 |
hy_whisper_local_cs,0.26486292350053875,0.076019400035926
|
13 |
phi_4_multimodal_instruct,1.7344522925894887,1.2856834920064666
|
14 |
+
whisper_large_v3,0.21394469053034837,0.028201904077600145
|
results_organized/wer/asr_vietnamese.csv
CHANGED
@@ -6,8 +6,9 @@ MERaLiON-AudioLLM-v2-9b-asr,0.1423883125132331,0.09499798648962901
|
|
6 |
Qwen2.5-Omni-3B,0.19648528477662502,0.17708681916408126
|
7 |
Qwen2.5-Omni-7B,0.18367562989625238,0.22730546937479085
|
8 |
SALMONN_7B,1.496294727927165,1.5460526688938172
|
|
|
9 |
cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct,0.1567859411391065,0.17136958408249153
|
10 |
cascade_whisper_large_v3_llama_3_8b_instruct,0.11676900275248782,0.15538061017872032
|
11 |
hy_whisper_local_cs,0.1681134871903451,0.1781020821398794
|
12 |
phi_4_multimodal_instruct,1.1070294304467498,2.5042567310800923
|
13 |
-
whisper_large_v3,0.
|
|
|
6 |
Qwen2.5-Omni-3B,0.19648528477662502,0.17708681916408126
|
7 |
Qwen2.5-Omni-7B,0.18367562989625238,0.22730546937479085
|
8 |
SALMONN_7B,1.496294727927165,1.5460526688938172
|
9 |
+
SeaLLMs-Audio-7B,0.25513444844378574,0.18886727279946003
|
10 |
cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct,0.1567859411391065,0.17136958408249153
|
11 |
cascade_whisper_large_v3_llama_3_8b_instruct,0.11676900275248782,0.15538061017872032
|
12 |
hy_whisper_local_cs,0.1681134871903451,0.1781020821398794
|
13 |
phi_4_multimodal_instruct,1.1070294304467498,2.5042567310800923
|
14 |
+
whisper_large_v3,0.13667160702943046,0.1602754255313115
|