Upload folder using huggingface_hub

results/bleu/st.csv

@@ -6,3 +6,5 @@ meralion_audiollm_v1_lora,1.983753597754947,0.24380649078367136,0.13427851707976
 whisper_large_v3_with_llama_3_8b_instruct,,,,,,
 mowe_audio,,,,,,
 qwen_audio_chat,4.121393814607146,15.307536203629063,0.030392249793261363,0.5089433695608178,10.019503192969147,0.018955183488794443
+meralion_audiollm_v1_mse,3.4633065594678443,4.974426298395461,0.19029087365857586,2.2174837053079424,1.1308873699909163,0.039170054539659147
+stage2_whisper3_fft_mlp100_gemma2_9b_lora,31.486376219193552,37.72055065660984,7.85766371365086,31.30757966907217,14.725914408637212,3.025625434709311

results/llama3_70b_judge/ac.csv

@@ -2,7 +2,9 @@ Model,audiocaps_test,wavcaps_test
 salmonn_7b,34.372727272727275,21.15606936416185
 wavllm_fairseq,5.5,6.901734104046243
 Qwen2-Audio-7B-Instruct,40.77727272727273,33.78034682080925
-meralion_audiollm_v1_lora,,28.080924855491332
+meralion_audiollm_v1_lora,38.972727272727276,28.080924855491332
 whisper_large_v3_with_llama_3_8b_instruct,2.4727272727272727,3.445086705202312
 mowe_audio,41.20909090909091,23.294797687861273
 qwen_audio_chat,43.82272727272727,29.248554913294797
+meralion_audiollm_v1_mse,4.2,5.144508670520231
+stage2_whisper3_fft_mlp100_gemma2_9b_lora,32.236363636363635,25.14450867052023

results/llama3_70b_judge/aqa.csv

@@ -6,3 +6,5 @@ meralion_audiollm_v1_lora,61.57510938259601,47.6038338658147,41.25
 whisper_large_v3_with_llama_3_8b_instruct,29.47134606841404,17.380191693290733,16.710526315789473
 mowe_audio,62.221235006663704,32.97124600638978,28.88157894736842
 qwen_audio_chat,58.196357174589075,47.98722044728435,38.68421052631579
+meralion_audiollm_v1_mse,,19.233226837060702,15.723684210526317
+stage2_whisper3_fft_mlp100_gemma2_9b_lora,55.303840544482256,38.65814696485623,34.53947368421053

results/llama3_70b_judge/ar.csv

@@ -6,3 +6,5 @@ meralion_audiollm_v1_lora,54.89946655724252
 whisper_large_v3_with_llama_3_8b_instruct,39.32704144439885
 mowe_audio,23.68485843249897
 qwen_audio_chat,45.699630693475584
+meralion_audiollm_v1_mse,62.83545342634387
+stage2_whisper3_fft_mlp100_gemma2_9b_lora,

results/llama3_70b_judge/si.csv

@@ -6,3 +6,5 @@ meralion_audiollm_v1_lora,65.0,71.39999999999999
 whisper_large_v3_with_llama_3_8b_instruct,63.0,70.8
 mowe_audio,16.0,19.8
 qwen_audio_chat,11.0,9.6
+meralion_audiollm_v1_mse,40.0,38.8
+stage2_whisper3_fft_mlp100_gemma2_9b_lora,63.6,69.2

results/llama3_70b_judge/sqa.csv

@@ -6,3 +6,5 @@ meralion_audiollm_v1_lora,81.17647058823529,53.51744186046511,66.30162586432442
 whisper_large_v3_with_llama_3_8b_instruct,82.99019607843137,64.94186046511628,83.81984675761541
 mowe_audio,76.5686274509804,,63.0349467389273
 qwen_audio_chat,76.12745098039215,57.47093023255814,60.65408334890675
+meralion_audiollm_v1_mse,82.05882352941175,,68.56288544197346
+stage2_whisper3_fft_mlp100_gemma2_9b_lora,77.84313725490196,53.11046511627907,

results/llama3_70b_judge_binary/er.csv

@@ -6,3 +6,5 @@ meralion_audiollm_v1_lora,38.822355289421154,52.95019157088122,44.67432950191571
 whisper_large_v3_with_llama_3_8b_instruct,34.4311377245509,43.86973180076628,33.25670498084291
 mowe_audio,23.55289421157685,46.09195402298851,30.07662835249042
 qwen_audio_chat,27.345309381237527,43.86973180076628,50.57471264367817
+meralion_audiollm_v1_mse,75.3493013972056,,79.88505747126436
+stage2_whisper3_fft_mlp100_gemma2_9b_lora,33.63273453093812,47.43295019157088,37.47126436781609

results/llama3_70b_judge_binary/gr.csv

@@ -6,3 +6,5 @@ meralion_audiollm_v1_lora,99.52810832991383,50.998003992015974
 whisper_large_v3_with_llama_3_8b_instruct,53.40582683627411,51.49700598802395
 mowe_audio,81.20640131308986,37.82435129740519
 qwen_audio_chat,70.55806319244974,51.796407185628745
+meralion_audiollm_v1_mse,47.250718096019696,
+stage2_whisper3_fft_mlp100_gemma2_9b_lora,85.06360279031597,48.60279441117765

results/llama3_70b_judge_binary/sqa.csv

@@ -6,3 +6,5 @@ meralion_audiollm_v1_lora,87.27432848965213,76.94720334553058
 whisper_large_v3_with_llama_3_8b_instruct,85.24878907970057,86.0951385258756
 mowe_audio,75.3412593571114,
 qwen_audio_chat,60.85424922941436,57.76267642446419
+meralion_audiollm_v1_mse,84.36811977102599,78.46314688970205
+stage2_whisper3_fft_mlp100_gemma2_9b_lora,83.92778511668868,80.39728175640354

results/meteor/ac.csv

@@ -6,3 +6,5 @@ meralion_audiollm_v1_lora,,
 whisper_large_v3_with_llama_3_8b_instruct,,
 mowe_audio,,
 qwen_audio_chat,,
+meralion_audiollm_v1_mse,,
+stage2_whisper3_fft_mlp100_gemma2_9b_lora,,

results/wer/asr.csv

@@ -2,7 +2,9 @@ Model,librispeech_test_clean,librispeech_test_other,common_voice_15_en_test,peop
 salmonn_7b,0.555834307918663,0.4180490665254317,0.3375363911542985,0.34335568653146914,0.14226764839341272,0.2687240831119165,0.3638199574803293,0.08558958163301515,0.18386810849140223,0.1581255622670608,0.6915082052617869
 wavllm_fairseq,0.02103218017882069,0.04798834811886432,0.14533325621300636,0.3792176325635977,0.15491778414546403,0.6447482518259942,0.6671766188447099,0.06621482559171073,0.4536784258110264,0.07762498393522684,0.4511070591299818
 Qwen2-Audio-7B-Instruct,0.03201041234390915,0.06073732195887794,0.11438872500819404,0.22323599266520214,0.11891779417331638,0.9865371146513101,0.9884312836905533,0.06387282211419042,0.953518879631271,0.0558572162961059,0.25686376660588695
-meralion_audiollm_v1_lora,0.023088240842041724,0.05005012578735317,0.0918438827513962,0.248582637787413,0.10348492939138457,0.15873920245466977,,0.042404456903587524,0.06544938840631094,0.04006554427451484,0.04100026048450117
+meralion_audiollm_v1_lora,0.023088240842041724,0.05005012578735317,0.0918438827513962,0.248582637787413,0.10348492939138457,0.15873920245466977,0.20147977375561024,0.042404456903587524,0.06544938840631094,0.04006554427451484,0.04100026048450117
 whisper_large_v3_with_llama_3_8b_instruct,0.01831591655034519,0.03714982881570734,0.09876543209876543,0.14540692118393275,0.09515429104337297,0.11773910240019567,0.15611126487402763,0.038146268762641496,0.04754476156709803,0.0464432592211798,0.2904141703568638
 mowe_audio,0.023258007318821442,0.05544101235175062,0.5609539784448686,0.6840078629561348,0.32259099649275863,0.32744494224633414,0.41765995821028074,0.2092544622263227,0.16780712639602907,,
 qwen_audio_chat,0.02246576376051609,0.04163277659031154,0.11645169375518152,0.30723956433545313,0.13316235845059654,0.38460383986837277,0.511809498898483,0.03995599872254356,0.45289842226555577,0.08434005911836524,0.5057046105756707
+meralion_audiollm_v1_mse,0.45899196438676576,0.5464089129324531,1.0056747707276945,0.811818861521855,0.4370689590337161,0.3729613900901602,0.4129325088937485,0.3558425889783897,0.24396383619925546,0.4923371032001028,1.869158635061214
+stage2_whisper3_fft_mlp100_gemma2_9b_lora,0.022031916097634587,0.04416743904515104,0.09438242684815651,0.25788482518786754,0.10128130883529002,0.15378928527754,0.20857550992683252,0.040701181647209116,0.058677539443361106,0.015052692455982522,0.044282365199270646

results/wer/cnasr.csv

@@ -6,3 +6,5 @@ meralion_audiollm_v1_lora,0.08340757458771357
 whisper_large_v3_with_llama_3_8b_instruct,
 mowe_audio,1.0174318465284788
 qwen_audio_chat,0.06602522222882144
+meralion_audiollm_v1_mse,2.1331096196868007
+stage2_whisper3_fft_mlp100_gemma2_9b_lora,0.08288293639009325