add transcripts for whisper-v3-turbo and first eval results

evaluation/cv_17/empty_stats/empty_stats_openai--whisper-large-v3-turbo_cv_17_devtest_ar_gender_male_masculine_female_feminine.json

@@ -0,0 +1,4 @@
+{
+  "reference": 0,
+  "transcription": 0
+}

evaluation/cv_17/empty_stats/empty_stats_openai--whisper-large-v3-turbo_cv_17_devtest_ca_gender_male_masculine_female_feminine.json

@@ -0,0 +1,4 @@
+{
+  "reference": 0,
+  "transcription": 0
+}

evaluation/cv_17/empty_stats/empty_stats_openai--whisper-large-v3-turbo_cv_17_devtest_cs_gender_male_masculine_female_feminine.json

@@ -0,0 +1,4 @@
+{
+  "reference": 0,
+  "transcription": 0
+}

evaluation/cv_17/empty_stats/empty_stats_openai--whisper-large-v3-turbo_cv_17_devtest_de_gender_male_masculine_female_feminine.json

@@ -0,0 +1,4 @@
+{
+  "reference": -1,
+  "transcription": 0
+}

evaluation/cv_17/empty_stats/empty_stats_openai--whisper-large-v3-turbo_cv_17_devtest_en_gender_male_masculine_female_feminine.json

@@ -0,0 +1,4 @@
+{
+  "reference": 0,
+  "transcription": 0
+}

evaluation/cv_17/empty_stats/empty_stats_openai--whisper-large-v3-turbo_cv_17_devtest_es_gender_male_masculine_female_feminine.json

@@ -0,0 +1,4 @@
+{
+  "reference": 0,
+  "transcription": 0
+}

evaluation/cv_17/empty_stats/empty_stats_openai--whisper-large-v3-turbo_cv_17_devtest_fi_gender_male_masculine_female_feminine.json

@@ -0,0 +1,4 @@
+{
+  "reference": 0,
+  "transcription": 0
+}

evaluation/cv_17/empty_stats/empty_stats_openai--whisper-large-v3-turbo_cv_17_devtest_fr_gender_male_masculine_female_feminine.json

@@ -0,0 +1,4 @@
+{
+  "reference": 0,
+  "transcription": 0
+}

evaluation/cv_17/empty_stats/empty_stats_openai--whisper-large-v3-turbo_cv_17_devtest_hu_gender_male_masculine_female_feminine.json

@@ -0,0 +1,4 @@
+{
+  "reference": 0,
+  "transcription": 0
+}

evaluation/cv_17/empty_stats/empty_stats_openai--whisper-large-v3-turbo_cv_17_devtest_it_gender_male_masculine_female_feminine.json

@@ -0,0 +1,4 @@
+{
+  "reference": 0,
+  "transcription": 0
+}

evaluation/cv_17/empty_stats/empty_stats_openai--whisper-large-v3-turbo_cv_17_devtest_ja_gender_male_masculine_female_feminine.json

@@ -0,0 +1,4 @@
+{
+  "reference": 0,
+  "transcription": 0
+}

evaluation/cv_17/empty_stats/empty_stats_openai--whisper-large-v3-turbo_cv_17_devtest_nl_gender_male_masculine_female_feminine.json

@@ -0,0 +1,4 @@
+{
+  "reference": 0,
+  "transcription": 0
+}

evaluation/cv_17/empty_stats/empty_stats_openai--whisper-large-v3-turbo_cv_17_devtest_pt_gender_male_masculine_female_feminine.json

@@ -0,0 +1,4 @@
+{
+  "reference": 0,
+  "transcription": 0
+}

evaluation/cv_17/empty_stats/empty_stats_openai--whisper-large-v3-turbo_cv_17_devtest_ro_gender_male_masculine_female_feminine.json

@@ -0,0 +1,4 @@
+{
+  "reference": 0,
+  "transcription": 0
+}

evaluation/cv_17/empty_stats/empty_stats_openai--whisper-large-v3-turbo_cv_17_devtest_ru_gender_male_masculine_female_feminine.json

@@ -0,0 +1,4 @@
+{
+  "reference": 0,
+  "transcription": 0
+}

evaluation/cv_17/empty_stats/empty_stats_openai--whisper-large-v3-turbo_cv_17_devtest_sk_gender_male_masculine_female_feminine.json

@@ -0,0 +1,4 @@
+{
+  "reference": 0,
+  "transcription": 0
+}

evaluation/cv_17/empty_stats/empty_stats_openai--whisper-large-v3-turbo_cv_17_devtest_sr_gender_male_masculine_female_feminine.json

@@ -0,0 +1,4 @@
+{
+  "reference": 0,
+  "transcription": 0
+}

evaluation/cv_17/empty_stats/empty_stats_openai--whisper-large-v3-turbo_cv_17_devtest_sw_gender_male_masculine_female_feminine.json

@@ -0,0 +1,4 @@
+{
+  "reference": 0,
+  "transcription": 0
+}

evaluation/cv_17/empty_stats/empty_stats_openai--whisper-large-v3-turbo_cv_17_devtest_yo_gender_male_masculine_female_feminine.json

@@ -0,0 +1,4 @@
+{
+  "reference": 0,
+  "transcription": 0
+}

evaluation/cv_17/empty_stats/empty_stats_openai--whisper-large-v3_cv_17_devtest_ar_gender_male_masculine_female_feminine.json

@@ -0,0 +1,4 @@
+{
+  "reference": 0,
+  "transcription": 0
+}

evaluation/cv_17/empty_stats/empty_stats_openai--whisper-large-v3_cv_17_devtest_ca_gender_male_masculine_female_feminine.json

@@ -0,0 +1,4 @@
+{
+  "reference": 0,
+  "transcription": 0
+}

evaluation/cv_17/empty_stats/empty_stats_openai--whisper-large-v3_cv_17_devtest_cs_gender_male_masculine_female_feminine.json

@@ -0,0 +1,4 @@
+{
+  "reference": 0,
+  "transcription": 0
+}

evaluation/cv_17/empty_stats/empty_stats_openai--whisper-large-v3_cv_17_devtest_de_gender_male_masculine_female_feminine.json

@@ -0,0 +1,4 @@
+{
+  "reference": -1,
+  "transcription": 0
+}

evaluation/cv_17/empty_stats/empty_stats_openai--whisper-large-v3_cv_17_devtest_en_gender_male_masculine_female_feminine.json

@@ -0,0 +1,4 @@
+{
+  "reference": 0,
+  "transcription": 0
+}

evaluation/cv_17/empty_stats/empty_stats_openai--whisper-large-v3_cv_17_devtest_es_gender_male_masculine_female_feminine.json

@@ -0,0 +1,4 @@
+{
+  "reference": 0,
+  "transcription": 0
+}

evaluation/cv_17/empty_stats/empty_stats_openai--whisper-large-v3_cv_17_devtest_fi_gender_male_masculine_female_feminine.json

@@ -0,0 +1,4 @@
+{
+  "reference": 0,
+  "transcription": 0
+}

evaluation/cv_17/empty_stats/empty_stats_openai--whisper-large-v3_cv_17_devtest_fr_gender_male_masculine_female_feminine.json

@@ -0,0 +1,4 @@
+{
+  "reference": 0,
+  "transcription": 0
+}

evaluation/cv_17/empty_stats/empty_stats_openai--whisper-large-v3_cv_17_devtest_hu_gender_male_masculine_female_feminine.json

@@ -0,0 +1,4 @@
+{
+  "reference": 0,
+  "transcription": 0
+}

evaluation/cv_17/empty_stats/empty_stats_openai--whisper-large-v3_cv_17_devtest_it_gender_male_masculine_female_feminine.json

@@ -0,0 +1,4 @@
+{
+  "reference": 0,
+  "transcription": 0
+}

evaluation/cv_17/empty_stats/empty_stats_openai--whisper-large-v3_cv_17_devtest_ja_gender_male_masculine_female_feminine.json

@@ -0,0 +1,4 @@
+{
+  "reference": 0,
+  "transcription": 0
+}

evaluation/cv_17/empty_stats/empty_stats_openai--whisper-large-v3_cv_17_devtest_nl_gender_male_masculine_female_feminine.json

@@ -0,0 +1,4 @@
+{
+  "reference": 0,
+  "transcription": 0
+}

evaluation/cv_17/empty_stats/empty_stats_openai--whisper-large-v3_cv_17_devtest_pt_gender_male_masculine_female_feminine.json

@@ -0,0 +1,4 @@
+{
+  "reference": 0,
+  "transcription": 0
+}

evaluation/cv_17/empty_stats/empty_stats_openai--whisper-large-v3_cv_17_devtest_ro_gender_male_masculine_female_feminine.json

@@ -0,0 +1,4 @@
+{
+  "reference": 0,
+  "transcription": 0
+}

evaluation/cv_17/empty_stats/empty_stats_openai--whisper-large-v3_cv_17_devtest_ru_gender_male_masculine_female_feminine.json

@@ -0,0 +1,4 @@
+{
+  "reference": 0,
+  "transcription": 0
+}

evaluation/cv_17/empty_stats/empty_stats_openai--whisper-large-v3_cv_17_devtest_sk_gender_male_masculine_female_feminine.json

@@ -0,0 +1,4 @@
+{
+  "reference": 0,
+  "transcription": 0
+}

evaluation/cv_17/empty_stats/empty_stats_openai--whisper-large-v3_cv_17_devtest_sr_gender_male_masculine_female_feminine.json

@@ -0,0 +1,4 @@
+{
+  "reference": 0,
+  "transcription": 0
+}

evaluation/cv_17/empty_stats/empty_stats_openai--whisper-large-v3_cv_17_devtest_sw_gender_male_masculine_female_feminine.json

@@ -0,0 +1,4 @@
+{
+  "reference": 0,
+  "transcription": 0
+}

evaluation/cv_17/empty_stats/empty_stats_openai--whisper-large-v3_cv_17_devtest_yo_gender_male_masculine_female_feminine.json

@@ -0,0 +1,4 @@
+{
+  "reference": 0,
+  "transcription": 0
+}

evaluation/cv_17/results_openai--whisper-large-v3-turbo_cv_17_devtest_ar_gender_male_masculine_female_feminine.json

@@ -0,0 +1,54 @@
+{
+  "model_id": "openai--whisper-large-v3-turbo",
+  "dataset_id": "cv_17",
+  "lang": "ar",
+  "split": "devtest",
+  "minority_group": "female_feminine",
+  "majority_group": "male_masculine",
+  "target_col": "gender",
+  "eval_metric": "wer",
+  "presample_wer": 0.6033332755842761,
+  "presample_cer": 0.15339151858258637,
+  "presample_maj_wer": 0.5535939261322194,
+  "presample_maj_cer": 0.14162965737421299,
+  "presample_min_wer": 0.6734036170395131,
+  "presample_min_cer": 0.17502554540762044,
+  "presample_diff_wer": 0.11980969090729376,
+  "presample_diff_cer": 0.033395888033407456,
+  "largest_group": "majority",
+  "overall_wer": 0.6124946274618865,
+  "overall_cer": 0.15694015830930683,
+  "overall_maj_wer": 0.5535939261322194,
+  "overall_maj_cer": 0.14162965737421299,
+  "overall_min_wer": 0.6734036170395131,
+  "overall_min_cer": 0.17502554540762044,
+  "overall_diff_wer": 0.11980969090729376,
+  "overall_diff_cer": 0.033395888033407456,
+  "overall_sample_size": 8008,
+  "mean_maj_wer": 0.5766525378551014,
+  "mean_min_wer": 0.7180874675568399,
+  "mean_maj_cer": 0.141803664295142,
+  "mean_min_cer": 0.19245266021050156,
+  "wer_diff_mean": 0.1414349297017386,
+  "wer_diff_std": 0.01213487074322525,
+  "wer_diff_rel_mean": 0.24567959451716423,
+  "wer_diff_rel_std": 0.02505822674731244,
+  "cer_diff_mean": 0.05064899591535958,
+  "cer_diff_std": 0.008633803338591867,
+  "cer_diff_rel_mean": 0.36070801701079486,
+  "cer_diff_rel_std": 0.07632953739940807,
+  "subsample_size": 4852,
+  "mean_subsample_wer": 0.656174315627421,
+  "mean_subsample_cer": 0.16850374729075918,
+  "1sided_ttest_wer_p": 0.0,
+  "1sided_ttest_wer_stat": 451.4484050927949,
+  "1sided_ttest_cer_p": 0.0,
+  "1sided_ttest_cer_stat": 224.2717035516984,
+  "2sided_ttest_wer_p": 0.0,
+  "2sided_ttest_wer_stat": 451.4484050927949,
+  "2sided_ttest_cer_p": 0.0,
+  "2sided_ttest_cer_stat": 224.2717035516984,
+  "minority_accept_percentile": 99,
+  "n_iterations": 1000,
+  "minority_perc_sampled": 0.4
+}

evaluation/cv_17/results_openai--whisper-large-v3-turbo_cv_17_devtest_ca_gender_male_masculine_female_feminine.json

@@ -0,0 +1,54 @@
+{
+  "model_id": "openai--whisper-large-v3-turbo",
+  "dataset_id": "cv_17",
+  "lang": "ca",
+  "split": "devtest",
+  "minority_group": "female_feminine",
+  "majority_group": "male_masculine",
+  "target_col": "gender",
+  "eval_metric": "wer",
+  "presample_wer": 0.1534435514383004,
+  "presample_cer": 0.06516574474098599,
+  "presample_maj_wer": 0.14197662061636557,
+  "presample_maj_cer": 0.05338177198102158,
+  "presample_min_wer": 0.12971189237212843,
+  "presample_min_cer": 0.04956476240308234,
+  "presample_diff_wer": -0.012264728244237133,
+  "presample_diff_cer": -0.0038170095779392413,
+  "largest_group": "majority",
+  "overall_wer": 0.13638729828214471,
+  "overall_cer": 0.05164522446683041,
+  "overall_maj_wer": 0.14197662061636557,
+  "overall_maj_cer": 0.05338177198102158,
+  "overall_min_wer": 0.12971189237212843,
+  "overall_min_cer": 0.04956476240308234,
+  "overall_diff_wer": -0.012264728244237133,
+  "overall_diff_cer": -0.0038170095779392413,
+  "overall_sample_size": 882,
+  "mean_maj_wer": 0.13627033525348242,
+  "mean_min_wer": 0.12328181856870117,
+  "mean_maj_cer": 0.05100113522782313,
+  "mean_min_cer": 0.04790251360914575,
+  "wer_diff_mean": -0.012988516684781239,
+  "wer_diff_std": 0.011426259532327917,
+  "wer_diff_rel_mean": -0.09172348389736727,
+  "wer_diff_rel_std": 0.08028070793009223,
+  "cer_diff_mean": -0.0030986216186773825,
+  "cer_diff_std": 0.006646098535619277,
+  "cer_diff_rel_mean": -0.052807513820233985,
+  "cer_diff_rel_std": 0.1288765562024992,
+  "subsample_size": 616,
+  "mean_subsample_wer": 0.1299316666181268,
+  "mean_subsample_cer": 0.04948940561641685,
+  "1sided_ttest_wer_p": 1.0,
+  "1sided_ttest_wer_stat": -43.928291052465624,
+  "1sided_ttest_cer_p": 1.0,
+  "1sided_ttest_cer_stat": -17.97848991940028,
+  "2sided_ttest_wer_p": 0.0,
+  "2sided_ttest_wer_stat": -43.928291052465624,
+  "2sided_ttest_cer_p": 1.0287216352771111e-68,
+  "2sided_ttest_cer_stat": -17.97848991940028,
+  "minority_accept_percentile": 99,
+  "n_iterations": 1000,
+  "minority_perc_sampled": 0.4
+}

evaluation/cv_17/results_openai--whisper-large-v3-turbo_cv_17_devtest_cs_gender_male_masculine_female_feminine.json

@@ -0,0 +1,54 @@
+{
+  "model_id": "openai--whisper-large-v3-turbo",
+  "dataset_id": "cv_17",
+  "lang": "cs",
+  "split": "devtest",
+  "minority_group": "female_feminine",
+  "majority_group": "male_masculine",
+  "target_col": "gender",
+  "eval_metric": "wer",
+  "presample_wer": 0.10850303007490496,
+  "presample_cer": 0.03032931347086193,
+  "presample_maj_wer": 0.10851043575525357,
+  "presample_maj_cer": 0.029759750033341588,
+  "presample_min_wer": 0.08328796951551443,
+  "presample_min_cer": 0.02134944957112774,
+  "presample_diff_wer": -0.025222466239739144,
+  "presample_diff_cer": -0.00841030046221385,
+  "largest_group": "majority",
+  "overall_wer": 0.10398601671744395,
+  "overall_cer": 0.028228671877921982,
+  "overall_maj_wer": 0.10851043575525357,
+  "overall_maj_cer": 0.029759750033341588,
+  "overall_min_wer": 0.08328796951551443,
+  "overall_min_cer": 0.02134944957112774,
+  "overall_diff_wer": -0.025222466239739144,
+  "overall_diff_cer": -0.00841030046221385,
+  "overall_sample_size": 11360,
+  "mean_maj_wer": 0.10500576563373178,
+  "mean_min_wer": 0.07155748025088551,
+  "mean_maj_cer": 0.028804291513903877,
+  "mean_min_cer": 0.018099469636072445,
+  "wer_diff_mean": -0.03344828538284625,
+  "wer_diff_std": 0.00553653289055918,
+  "wer_diff_rel_mean": -0.3172129654763564,
+  "wer_diff_rel_std": 0.041845900664913165,
+  "cer_diff_mean": -0.01070482187783143,
+  "cer_diff_std": 0.00221295847708785,
+  "cer_diff_rel_mean": -0.3689325319054002,
+  "cer_diff_rel_std": 0.05767286096712889,
+  "subsample_size": 1912,
+  "mean_subsample_wer": 0.08818898663931181,
+  "mean_subsample_cer": 0.023398180941175545,
+  "1sided_ttest_wer_p": 1.0,
+  "1sided_ttest_wer_stat": -234.7013216129715,
+  "1sided_ttest_cer_p": 1.0,
+  "1sided_ttest_cer_stat": -190.10202342244403,
+  "2sided_ttest_wer_p": 0.0,
+  "2sided_ttest_wer_stat": -234.7013216129715,
+  "2sided_ttest_cer_p": 0.0,
+  "2sided_ttest_cer_stat": -190.10202342244403,
+  "minority_accept_percentile": 99,
+  "n_iterations": 1000,
+  "minority_perc_sampled": 0.4
+}

evaluation/cv_17/results_openai--whisper-large-v3-turbo_cv_17_devtest_de_gender_male_masculine_female_feminine.json

@@ -0,0 +1,54 @@
+{
+  "model_id": "openai--whisper-large-v3-turbo",
+  "dataset_id": "cv_17",
+  "lang": "de",
+  "split": "devtest",
+  "minority_group": "female_feminine",
+  "majority_group": "male_masculine",
+  "target_col": "gender",
+  "eval_metric": "wer",
+  "presample_wer": 0.061284380444160194,
+  "presample_cer": 0.01812794653534037,
+  "presample_maj_wer": 0.06701068927386657,
+  "presample_maj_cer": 0.021290558979715726,
+  "presample_min_wer": 0.08793410507569012,
+  "presample_min_cer": 0.026570599550092916,
+  "presample_diff_wer": 0.020923415801823553,
+  "presample_diff_cer": 0.00528004057037719,
+  "largest_group": "majority",
+  "overall_wer": 0.06998292328125988,
+  "overall_cer": 0.02204904366722242,
+  "overall_maj_wer": 0.06701068927386657,
+  "overall_maj_cer": 0.021290558979715726,
+  "overall_min_wer": 0.08793410507569012,
+  "overall_min_cer": 0.026570599550092916,
+  "overall_diff_wer": 0.020923415801823553,
+  "overall_diff_cer": 0.00528004057037719,
+  "overall_sample_size": 3006,
+  "mean_maj_wer": 0.06665010415157135,
+  "mean_min_wer": 0.08744226395624011,
+  "mean_maj_cer": 0.020821314253914595,
+  "mean_min_cer": 0.026054112930103807,
+  "wer_diff_mean": 0.020792159804668756,
+  "wer_diff_std": 0.011505161563902406,
+  "wer_diff_rel_mean": 0.3357636531174909,
+  "wer_diff_rel_std": 0.21469342545738,
+  "cer_diff_mean": 0.0052327986761892134,
+  "cer_diff_std": 0.005032036958802708,
+  "cer_diff_rel_mean": 0.2952141061264395,
+  "cer_diff_rel_std": 0.29423463353176493,
+  "subsample_size": 390,
+  "mean_subsample_wer": 0.07714336236354939,
+  "mean_subsample_cer": 0.023478924223182315,
+  "1sided_ttest_wer_p": 0.0,
+  "1sided_ttest_wer_stat": 72.39863274538659,
+  "1sided_ttest_cer_p": 2.730071185855925e-295,
+  "1sided_ttest_cer_stat": 41.25027618171636,
+  "2sided_ttest_wer_p": 0.0,
+  "2sided_ttest_wer_stat": 72.39863274538659,
+  "2sided_ttest_cer_p": 5.46014237171185e-295,
+  "2sided_ttest_cer_stat": 41.25027618171636,
+  "minority_accept_percentile": 99,
+  "n_iterations": 1000,
+  "minority_perc_sampled": 0.4
+}

evaluation/cv_17/results_openai--whisper-large-v3-turbo_cv_17_devtest_en_gender_male_masculine_female_feminine.json

@@ -0,0 +1,54 @@
+{
+  "model_id": "openai--whisper-large-v3-turbo",
+  "dataset_id": "cv_17",
+  "lang": "en",
+  "split": "devtest",
+  "minority_group": "female_feminine",
+  "majority_group": "male_masculine",
+  "target_col": "gender",
+  "eval_metric": "wer",
+  "presample_wer": 0.10215818694590938,
+  "presample_cer": 0.04847400219958553,
+  "presample_maj_wer": 0.1318213128555757,
+  "presample_maj_cer": 0.06562171755586946,
+  "presample_min_wer": 0.09281401681070202,
+  "presample_min_cer": 0.04341713182524028,
+  "presample_diff_wer": -0.03900729604487367,
+  "presample_diff_cer": -0.02220458573062918,
+  "largest_group": "majority",
+  "overall_wer": 0.12490554949206616,
+  "overall_cer": 0.061681177986489207,
+  "overall_maj_wer": 0.1318213128555757,
+  "overall_maj_cer": 0.06562171755586946,
+  "overall_min_wer": 0.09281401681070202,
+  "overall_min_cer": 0.04341713182524028,
+  "overall_diff_wer": -0.03900729604487367,
+  "overall_diff_cer": -0.02220458573062918,
+  "overall_sample_size": 4086,
+  "mean_maj_wer": 0.13134657863148078,
+  "mean_min_wer": 0.09239075572851706,
+  "mean_maj_cer": 0.06378729830032513,
+  "mean_min_cer": 0.04207321121354682,
+  "wer_diff_mean": -0.0389558229029637,
+  "wer_diff_std": 0.052956929482409744,
+  "wer_diff_rel_mean": -0.2174266873794035,
+  "wer_diff_rel_std": 0.21131117053031698,
+  "cer_diff_mean": -0.021714087086778324,
+  "cer_diff_std": 0.034477678489026255,
+  "cer_diff_rel_mean": -0.22316835480312217,
+  "cer_diff_rel_std": 0.23934427605175757,
+  "subsample_size": 692,
+  "mean_subsample_wer": 0.11177063824557328,
+  "mean_subsample_cer": 0.05288689610320853,
+  "1sided_ttest_wer_p": 1.0,
+  "1sided_ttest_wer_stat": -28.477299042186342,
+  "1sided_ttest_cer_p": 1.0,
+  "1sided_ttest_cer_stat": -24.411651006524465,
+  "2sided_ttest_wer_p": 4.377702079672933e-158,
+  "2sided_ttest_wer_stat": -28.477299042186342,
+  "2sided_ttest_cer_p": 3.34922820343185e-120,
+  "2sided_ttest_cer_stat": -24.411651006524465,
+  "minority_accept_percentile": 99,
+  "n_iterations": 1000,
+  "minority_perc_sampled": 0.4
+}

evaluation/cv_17/results_openai--whisper-large-v3-turbo_cv_17_devtest_es_gender_male_masculine_female_feminine.json

@@ -0,0 +1,54 @@
+{
+  "model_id": "openai--whisper-large-v3-turbo",
+  "dataset_id": "cv_17",
+  "lang": "es",
+  "split": "devtest",
+  "minority_group": "female_feminine",
+  "majority_group": "male_masculine",
+  "target_col": "gender",
+  "eval_metric": "wer",
+  "presample_wer": 0.048534536026652905,
+  "presample_cer": 0.01716668425602862,
+  "presample_maj_wer": 0.052762810353935555,
+  "presample_maj_cer": 0.018984769061648362,
+  "presample_min_wer": 0.04583487141626676,
+  "presample_min_cer": 0.015624195359088709,
+  "presample_diff_wer": -0.006927938937668791,
+  "presample_diff_cer": -0.003360573702559653,
+  "largest_group": "majority",
+  "overall_wer": 0.0509916796426493,
+  "overall_cer": 0.01812612498815802,
+  "overall_maj_wer": 0.052762810353935555,
+  "overall_maj_cer": 0.018984769061648362,
+  "overall_min_wer": 0.04583487141626676,
+  "overall_min_cer": 0.015624195359088709,
+  "overall_diff_wer": -0.006927938937668791,
+  "overall_diff_cer": -0.003360573702559653,
+  "overall_sample_size": 4827,
+  "mean_maj_wer": 0.052560865352823956,
+  "mean_min_wer": 0.04531754541582075,
+  "mean_maj_cer": 0.018798060972108482,
+  "mean_min_cer": 0.01547102704851853,
+  "wer_diff_mean": -0.007243319937003208,
+  "wer_diff_std": 0.01305171694847401,
+  "wer_diff_rel_mean": -0.09994143614398832,
+  "wer_diff_rel_std": 0.16655624099769875,
+  "cer_diff_mean": -0.0033270339235899533,
+  "cer_diff_std": 0.0064280888029596955,
+  "cer_diff_rel_mean": -0.11380600256263255,
+  "cer_diff_rel_std": 0.20851033876065903,
+  "subsample_size": 1300,
+  "mean_subsample_wer": 0.04890682076521527,
+  "mean_subsample_cer": 0.01712211841087596,
+  "1sided_ttest_wer_p": 1.0,
+  "1sided_ttest_wer_stat": -21.407392510010148,
+  "1sided_ttest_cer_p": 1.0,
+  "1sided_ttest_cer_stat": -20.040439413328613,
+  "2sided_ttest_wer_p": 9.949404729596734e-95,
+  "2sided_ttest_wer_stat": -21.407392510010148,
+  "2sided_ttest_cer_p": 6.065282892411041e-84,
+  "2sided_ttest_cer_stat": -20.040439413328613,
+  "minority_accept_percentile": 99,
+  "n_iterations": 1000,
+  "minority_perc_sampled": 0.4
+}

evaluation/cv_17/results_openai--whisper-large-v3-turbo_cv_17_devtest_fi_gender_male_masculine_female_feminine.json

@@ -0,0 +1,54 @@
+{
+  "model_id": "openai--whisper-large-v3-turbo",
+  "dataset_id": "cv_17",
+  "lang": "fi",
+  "split": "devtest",
+  "minority_group": "female_feminine",
+  "majority_group": "male_masculine",
+  "target_col": "gender",
+  "eval_metric": "wer",
+  "presample_wer": 0.12855798108386748,
+  "presample_cer": 0.0305606723463044,
+  "presample_maj_wer": 0.12694969527941327,
+  "presample_maj_cer": 0.024556307626712147,
+  "presample_min_wer": 0.1799265605875153,
+  "presample_min_cer": 0.04046948855488874,
+  "presample_diff_wer": 0.05297686530810203,
+  "presample_diff_cer": 0.01591318092817659,
+  "largest_group": "majority",
+  "overall_wer": 0.13765248928453677,
+  "overall_cer": 0.027747893609995633,
+  "overall_maj_wer": 0.12694969527941327,
+  "overall_maj_cer": 0.024556307626712147,
+  "overall_min_wer": 0.1799265605875153,
+  "overall_min_cer": 0.04046948855488874,
+  "overall_diff_wer": 0.05297686530810203,
+  "overall_diff_cer": 0.01591318092817659,
+  "overall_sample_size": 1526,
+  "mean_maj_wer": 0.13268496504061597,
+  "mean_min_wer": 0.21484684761805364,
+  "mean_maj_cer": 0.02714767961728714,
+  "mean_min_cer": 0.04949623595635918,
+  "wer_diff_mean": 0.08216188257743764,
+  "wer_diff_std": 0.019538185134450874,
+  "wer_diff_rel_mean": 0.6368974354179266,
+  "wer_diff_rel_std": 0.20196882898717988,
+  "cer_diff_mean": 0.022348556339072036,
+  "cer_diff_std": 0.006066734390514327,
+  "cer_diff_rel_mean": 0.875533333416221,
+  "cer_diff_rel_std": 0.34921147179240036,
+  "subsample_size": 310,
+  "mean_subsample_wer": 0.1737034480425112,
+  "mean_subsample_cer": 0.038249736803159524,
+  "1sided_ttest_wer_p": 0.0,
+  "1sided_ttest_wer_stat": 162.21715056251696,
+  "1sided_ttest_cer_p": 0.0,
+  "1sided_ttest_cer_stat": 143.28784672869767,
+  "2sided_ttest_wer_p": 0.0,
+  "2sided_ttest_wer_stat": 162.21715056251696,
+  "2sided_ttest_cer_p": 0.0,
+  "2sided_ttest_cer_stat": 143.28784672869767,
+  "minority_accept_percentile": 99,
+  "n_iterations": 1000,
+  "minority_perc_sampled": 0.4
+}

evaluation/cv_17/results_openai--whisper-large-v3-turbo_cv_17_devtest_fr_gender_male_masculine_female_feminine.json

@@ -0,0 +1,54 @@
+{
+  "model_id": "openai--whisper-large-v3-turbo",
+  "dataset_id": "cv_17",
+  "lang": "fr",
+  "split": "devtest",
+  "minority_group": "female_feminine",
+  "majority_group": "male_masculine",
+  "target_col": "gender",
+  "eval_metric": "wer",
+  "presample_wer": 0.10822041660163095,
+  "presample_cer": 0.046328094452938,
+  "presample_maj_wer": 0.12443611411144291,
+  "presample_maj_cer": 0.05957820934973008,
+  "presample_min_wer": 0.1310910387449489,
+  "presample_min_cer": 0.059425414827059486,
+  "presample_diff_wer": 0.00665492463350599,
+  "presample_diff_cer": -0.00015279452267059207,
+  "largest_group": "majority",
+  "overall_wer": 0.12583976737190414,
+  "overall_cer": 0.059545950530784625,
+  "overall_maj_wer": 0.12443611411144291,
+  "overall_maj_cer": 0.05957820934973008,
+  "overall_min_wer": 0.1310910387449489,
+  "overall_min_cer": 0.059425414827059486,
+  "overall_diff_wer": 0.00665492463350599,
+  "overall_diff_cer": -0.00015279452267059207,
+  "overall_sample_size": 3092,
+  "mean_maj_wer": 0.12034636871179115,
+  "mean_min_wer": 0.12696259076476243,
+  "mean_maj_cer": 0.05759678615420773,
+  "mean_min_cer": 0.05762560675265852,
+  "wer_diff_mean": 0.006616222052971299,
+  "wer_diff_std": 0.01276294409225713,
+  "wer_diff_rel_mean": 0.06217136456960209,
+  "wer_diff_rel_std": 0.11134450953620606,
+  "cer_diff_mean": 2.882059845077849e-05,
+  "cer_diff_std": 0.008468554996959,
+  "cer_diff_rel_mean": 0.014893477211854692,
+  "cer_diff_rel_std": 0.15319680185915643,
+  "subsample_size": 638,
+  "mean_subsample_wer": 0.12370261239034429,
+  "mean_subsample_cer": 0.057609220640056226,
+  "1sided_ttest_wer_p": 9.254453835860342e-85,
+  "1sided_ttest_wer_stat": 20.107348855297175,
+  "1sided_ttest_cer_p": 0.44775280973386794,
+  "1sided_ttest_cer_stat": 0.1313520658536847,
+  "2sided_ttest_wer_p": 1.8508907671720684e-84,
+  "2sided_ttest_wer_stat": 20.107348855297175,
+  "2sided_ttest_cer_p": 0.8955056194677359,
+  "2sided_ttest_cer_stat": 0.1313520658536847,
+  "minority_accept_percentile": 99,
+  "n_iterations": 1000,
+  "minority_perc_sampled": 0.4
+}

evaluation/cv_17/results_openai--whisper-large-v3-turbo_cv_17_devtest_hu_gender_male_masculine_female_feminine.json

@@ -0,0 +1,54 @@
+{
+  "model_id": "openai--whisper-large-v3-turbo",
+  "dataset_id": "cv_17",
+  "lang": "hu",
+  "split": "devtest",
+  "minority_group": "female_feminine",
+  "majority_group": "male_masculine",
+  "target_col": "gender",
+  "eval_metric": "wer",
+  "presample_wer": 0.1481324203337049,
+  "presample_cer": 0.03235697811686129,
+  "presample_maj_wer": 0.16809938439354444,
+  "presample_maj_cer": 0.03820401732965734,
+  "presample_min_wer": 0.13685463200292933,
+  "presample_min_cer": 0.028795435788869353,
+  "presample_diff_wer": -0.03124475239061511,
+  "presample_diff_cer": -0.00940858154078799,
+  "largest_group": "minority",
+  "overall_wer": 0.14927990119318058,
+  "overall_cer": 0.03252681312393192,
+  "overall_maj_wer": 0.16809938439354444,
+  "overall_maj_cer": 0.03820401732965734,
+  "overall_min_wer": 0.13685463200292933,
+  "overall_min_cer": 0.028795435788869353,
+  "overall_diff_wer": -0.03124475239061511,
+  "overall_diff_cer": -0.00940858154078799,
+  "overall_sample_size": 4882,
+  "mean_maj_wer": 0.15493889508008146,
+  "mean_min_wer": 0.13639694918261136,
+  "mean_maj_cer": 0.03404736918832601,
+  "mean_min_cer": 0.029161304158297846,
+  "wer_diff_mean": -0.018541945897470095,
+  "wer_diff_std": 0.0061327059451669755,
+  "wer_diff_rel_mean": -0.1186958520168174,
+  "wer_diff_rel_std": 0.03492822931427987,
+  "cer_diff_mean": -0.004886065030028163,
+  "cer_diff_std": 0.0027326511427123545,
+  "cer_diff_rel_mean": -0.141843395220745,
+  "cer_diff_rel_std": 0.07608326544748403,
+  "subsample_size": 3904,
+  "mean_subsample_wer": 0.1456068207213744,
+  "mean_subsample_cer": 0.03157913876801826,
+  "1sided_ttest_wer_p": 1.0,
+  "1sided_ttest_wer_stat": -117.29414452650263,
+  "1sided_ttest_cer_p": 1.0,
+  "1sided_ttest_cer_stat": -68.73612693559616,
+  "2sided_ttest_wer_p": 0.0,
+  "2sided_ttest_wer_stat": -117.29414452650263,
+  "2sided_ttest_cer_p": 0.0,
+  "2sided_ttest_cer_stat": -68.73612693559616,
+  "minority_accept_percentile": 99,
+  "n_iterations": 1000,
+  "minority_perc_sampled": 0.4
+}

evaluation/cv_17/results_openai--whisper-large-v3-turbo_cv_17_devtest_it_gender_male_masculine_female_feminine.json

@@ -0,0 +1,54 @@
+{
+  "model_id": "openai--whisper-large-v3-turbo",
+  "dataset_id": "cv_17",
+  "lang": "it",
+  "split": "devtest",
+  "minority_group": "female_feminine",
+  "majority_group": "male_masculine",
+  "target_col": "gender",
+  "eval_metric": "wer",
+  "presample_wer": 0.06071115376342307,
+  "presample_cer": 0.01917170556382645,
+  "presample_maj_wer": 0.05953713077920463,
+  "presample_maj_cer": 0.018339886058085984,
+  "presample_min_wer": 0.056476948929641266,
+  "presample_min_cer": 0.01734276055801552,
+  "presample_diff_wer": -0.003060181849563366,
+  "presample_diff_cer": -0.0009971255000704622,
+  "largest_group": "majority",
+  "overall_wer": 0.058938675201463106,
+  "overall_cer": 0.01814434241056214,
+  "overall_maj_wer": 0.05953713077920463,
+  "overall_maj_cer": 0.018339886058085984,
+  "overall_min_wer": 0.056476948929641266,
+  "overall_min_cer": 0.01734276055801552,
+  "overall_diff_wer": -0.003060181849563366,
+  "overall_diff_cer": -0.0009971255000704622,
+  "overall_sample_size": 5679,
+  "mean_maj_wer": 0.05778341638794547,
+  "mean_min_wer": 0.05514554967407353,
+  "mean_maj_cer": 0.017683925851908838,
+  "mean_min_cer": 0.01657518963626749,
+  "wer_diff_mean": -0.0026378667138719464,
+  "wer_diff_std": 0.005253653473156076,
+  "wer_diff_rel_mean": -0.040611190760584476,
+  "wer_diff_rel_std": 0.08926739689956975,
+  "cer_diff_mean": -0.001108736215641348,
+  "cer_diff_std": 0.0020572464085038948,
+  "cer_diff_rel_mean": -0.05466140291765295,
+  "cer_diff_rel_std": 0.11338128773904675,
+  "subsample_size": 1088,
+  "mean_subsample_wer": 0.05645167804493086,
+  "mean_subsample_cer": 0.017123158047116754,
+  "1sided_ttest_wer_p": 1.0,
+  "1sided_ttest_wer_stat": -19.344434783571867,
+  "1sided_ttest_cer_p": 1.0,
+  "1sided_ttest_cer_stat": -20.762763127545348,
+  "2sided_ttest_wer_p": 1.166523910953759e-78,
+  "2sided_ttest_wer_stat": -19.344434783571867,
+  "2sided_ttest_cer_p": 1.41192942864332e-89,
+  "2sided_ttest_cer_stat": -20.762763127545348,
+  "minority_accept_percentile": 99,
+  "n_iterations": 1000,
+  "minority_perc_sampled": 0.4
+}

evaluation/cv_17/results_openai--whisper-large-v3-turbo_cv_17_devtest_ja_gender_male_masculine_female_feminine.json

@@ -0,0 +1,54 @@
+{
+  "model_id": "openai--whisper-large-v3-turbo",
+  "dataset_id": "cv_17",
+  "lang": "ja",
+  "split": "devtest",
+  "minority_group": "female_feminine",
+  "majority_group": "male_masculine",
+  "target_col": "gender",
+  "eval_metric": "cer",
+  "presample_wer": 0.8457599864590386,
+  "presample_cer": 0.18463034318844754,
+  "presample_maj_wer": 0.8771096513390602,
+  "presample_maj_cer": 0.21260334171132292,
+  "presample_min_wer": 0.8426568038218524,
+  "presample_min_cer": 0.18185163830857368,
+  "presample_diff_wer": -0.034452847517207785,
+  "presample_diff_cer": -0.030751703402749242,
+  "largest_group": "majority",
+  "overall_wer": 0.86346435546875,
+  "overall_cer": 0.2005537423271598,
+  "overall_maj_wer": 0.8771096513390602,
+  "overall_maj_cer": 0.21260334171132292,
+  "overall_min_wer": 0.8426568038218524,
+  "overall_min_cer": 0.18185163830857368,
+  "overall_diff_wer": -0.034452847517207785,
+  "overall_diff_cer": -0.030751703402749242,
+  "overall_sample_size": 5310,
+  "mean_maj_wer": 0.8440851036592515,
+  "mean_min_wer": 0.8295883944006175,
+  "mean_maj_cer": 0.18342722881388507,
+  "mean_min_cer": 0.1752868519321126,
+  "wer_diff_mean": -0.014496709258634183,
+  "wer_diff_std": 0.028159313235585282,
+  "wer_diff_rel_mean": -0.016215899514923655,
+  "wer_diff_rel_std": 0.03092511418705418,
+  "cer_diff_mean": -0.008140376881772427,
+  "cer_diff_std": 0.01896153717922117,
+  "cer_diff_rel_mean": -0.03639116094010804,
+  "cer_diff_rel_std": 0.09850640493596836,
+  "subsample_size": 2446,
+  "mean_subsample_wer": 0.8363175986729127,
+  "mean_subsample_cer": 0.17910309921899997,
+  "1sided_ttest_wer_p": 1.0,
+  "1sided_ttest_wer_stat": -19.87020412107124,
+  "1sided_ttest_cer_p": 1.0,
+  "1sided_ttest_cer_stat": -16.410718093234316,
+  "2sided_ttest_wer_p": 1.2259000432834166e-82,
+  "2sided_ttest_wer_stat": -19.87020412107124,
+  "2sided_ttest_cer_p": 5.0444633527233204e-58,
+  "2sided_ttest_cer_stat": -16.410718093234316,
+  "minority_accept_percentile": 99,
+  "n_iterations": 1000,
+  "minority_perc_sampled": 0.4
+}

evaluation/cv_17/results_openai--whisper-large-v3-turbo_cv_17_devtest_nl_gender_male_masculine_female_feminine.json

@@ -0,0 +1,54 @@
+{
+  "model_id": "openai--whisper-large-v3-turbo",
+  "dataset_id": "cv_17",
+  "lang": "nl",
+  "split": "devtest",
+  "minority_group": "female_feminine",
+  "majority_group": "male_masculine",
+  "target_col": "gender",
+  "eval_metric": "wer",
+  "presample_wer": 0.0491471702423331,
+  "presample_cer": 0.016639717794658233,
+  "presample_maj_wer": 0.046926375602541044,
+  "presample_maj_cer": 0.015173360917004877,
+  "presample_min_wer": 0.04389305635699173,
+  "presample_min_cer": 0.014622227553681127,
+  "presample_diff_wer": -0.0030333192455493166,
+  "presample_diff_cer": -0.0005511333633237501,
+  "largest_group": "majority",
+  "overall_wer": 0.046269129463541496,
+  "overall_cer": 0.01505342068237777,
+  "overall_maj_wer": 0.046926375602541044,
+  "overall_maj_cer": 0.015173360917004877,
+  "overall_min_wer": 0.04389305635699173,
+  "overall_min_cer": 0.014622227553681127,
+  "overall_diff_wer": -0.0030333192455493166,
+  "overall_diff_cer": -0.0005511333633237501,
+  "overall_sample_size": 10102,
+  "mean_maj_wer": 0.03876061503645189,
+  "mean_min_wer": 0.03552316215792518,
+  "mean_maj_cer": 0.011639003665571527,
+  "mean_min_cer": 0.010916711640197835,
+  "wer_diff_mean": -0.0032374528785267085,
+  "wer_diff_std": 0.0027553925783330266,
+  "wer_diff_rel_mean": -0.0804274240534108,
+  "wer_diff_rel_std": 0.06794706317404613,
+  "cer_diff_mean": -0.000722292025373692,
+  "cer_diff_std": 0.00104558077338231,
+  "cer_diff_rel_mean": -0.05725397689259224,
+  "cer_diff_rel_std": 0.08766551924648483,
+  "subsample_size": 2182,
+  "mean_subsample_wer": 0.037130971052139285,
+  "mean_subsample_cer": 0.011276216825357883,
+  "1sided_ttest_wer_p": 1.0,
+  "1sided_ttest_wer_stat": -45.512746148959565,
+  "1sided_ttest_cer_p": 1.0,
+  "1sided_ttest_cer_stat": -26.729506633352624,
+  "2sided_ttest_wer_p": 0.0,
+  "2sided_ttest_wer_stat": -45.512746148959565,
+  "2sided_ttest_cer_p": 2.3290414007350592e-141,
+  "2sided_ttest_cer_stat": -26.729506633352624,
+  "minority_accept_percentile": 99,
+  "n_iterations": 1000,
+  "minority_perc_sampled": 0.4
+}