jonatasgrosman/exp_w2v2r_fr_xls-r_age_teens-8_sixties-2_s507

.gitattributes

@@ -32,3 +32,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+cv7_transcriptions.json filter=lfs diff=lfs merge=lfs -text

README.md

@@ -0,0 +1,17 @@
+---
+language:
+- fr
+license: apache-2.0
+tags:
+- automatic-speech-recognition
+- fr
+datasets:
+- mozilla-foundation/common_voice_7_0
+---
+# exp_w2v2r_fr_xls-r_age_teens-8_sixties-2_s507
+
+Fine-tuned [facebook/wav2vec2-xls-r-300m](https://huggingface.co/facebook/wav2vec2-xls-r-300m) for speech recognition using the train split of [Common Voice 7.0 (fr)](https://huggingface.co/datasets/mozilla-foundation/common_voice_7_0).
+When using this model, make sure that your speech input is sampled at 16kHz.
+
+This model has been fine-tuned by the [HuggingSound](https://github.com/jonatasgrosman/huggingsound) tool.
+    

config.json

@@ -0,0 +1,108 @@
+{
+  "_name_or_path": "facebook/wav2vec2-xls-r-300m",
+  "activation_dropout": 0.05,
+  "adapter_kernel_size": 3,
+  "adapter_stride": 2,
+  "add_adapter": false,
+  "apply_spec_augment": true,
+  "architectures": [
+    "Wav2Vec2ForCTC"
+  ],
+  "attention_dropout": 0.05,
+  "bos_token_id": 1,
+  "classifier_proj_size": 256,
+  "codevector_dim": 768,
+  "contrastive_logits_temperature": 0.1,
+  "conv_bias": true,
+  "conv_dim": [
+    512,
+    512,
+    512,
+    512,
+    512,
+    512,
+    512
+  ],
+  "conv_kernel": [
+    10,
+    3,
+    3,
+    3,
+    3,
+    2,
+    2
+  ],
+  "conv_stride": [
+    5,
+    2,
+    2,
+    2,
+    2,
+    2,
+    2
+  ],
+  "ctc_loss_reduction": "mean",
+  "ctc_zero_infinity": false,
+  "diversity_loss_weight": 0.1,
+  "do_stable_layer_norm": true,
+  "eos_token_id": 2,
+  "feat_extract_activation": "gelu",
+  "feat_extract_dropout": 0.0,
+  "feat_extract_norm": "layer",
+  "feat_proj_dropout": 0.05,
+  "feat_quantizer_dropout": 0.0,
+  "final_dropout": 0.05,
+  "gradient_checkpointing": false,
+  "hidden_act": "gelu",
+  "hidden_dropout": 0.05,
+  "hidden_size": 1024,
+  "initializer_range": 0.02,
+  "intermediate_size": 4096,
+  "layer_norm_eps": 1e-05,
+  "layerdrop": 0.05,
+  "mask_feature_length": 10,
+  "mask_feature_min_masks": 0,
+  "mask_feature_prob": 0.0,
+  "mask_time_length": 10,
+  "mask_time_min_masks": 2,
+  "mask_time_prob": 0.05,
+  "model_type": "wav2vec2",
+  "num_adapter_layers": 3,
+  "num_attention_heads": 16,
+  "num_codevector_groups": 2,
+  "num_codevectors_per_group": 320,
+  "num_conv_pos_embedding_groups": 16,
+  "num_conv_pos_embeddings": 128,
+  "num_feat_extract_layers": 7,
+  "num_hidden_layers": 24,
+  "num_negatives": 100,
+  "output_hidden_size": 1024,
+  "pad_token_id": 0,
+  "proj_codevector_dim": 768,
+  "tdnn_dilation": [
+    1,
+    2,
+    3,
+    1,
+    1
+  ],
+  "tdnn_dim": [
+    512,
+    512,
+    512,
+    512,
+    1500
+  ],
+  "tdnn_kernel": [
+    5,
+    3,
+    3,
+    1,
+    1
+  ],
+  "torch_dtype": "float32",
+  "transformers_version": "4.18.0",
+  "use_weighted_layer_sum": false,
+  "vocab_size": 57,
+  "xvector_output_dim": 512
+}

cv7_evaluation.json

@@ -0,0 +1,444 @@
+{
+  "validation": [
+    {
+      "subset_type": "all",
+      "subset": "all",
+      "cer": 0.07515291355647799,
+      "wer": 0.26328602738675644
+    },
+    {
+      "subset_type": "age",
+      "subset": "",
+      "cer": 0.07328745784009422,
+      "wer": 0.25824064396362106
+    },
+    {
+      "subset_type": "age",
+      "subset": "twenties",
+      "cer": 0.08650668199453944,
+      "wer": 0.2898194637325072
+    },
+    {
+      "subset_type": "age",
+      "subset": "fourties",
+      "cer": 0.08338206057771021,
+      "wer": 0.29041985618186034
+    },
+    {
+      "subset_type": "age",
+      "subset": "thirties",
+      "cer": 0.0834562152373561,
+      "wer": 0.28666050523721504
+    },
+    {
+      "subset_type": "age",
+      "subset": "fifties",
+      "cer": 0.08624553830829253,
+      "wer": 0.2807017543859649
+    },
+    {
+      "subset_type": "age",
+      "subset": "sixties",
+      "cer": 0.07100426305122651,
+      "wer": 0.2538527397260274
+    },
+    {
+      "subset_type": "age",
+      "subset": "seventies",
+      "cer": 0.07525163764179581,
+      "wer": 0.2796934865900383
+    },
+    {
+      "subset_type": "age",
+      "subset": "teens",
+      "cer": 0.08793906584413949,
+      "wer": 0.2998165137614679
+    },
+    {
+      "subset_type": "age",
+      "subset": "eighties",
+      "cer": 0.08,
+      "wer": 0.27586206896551724
+    },
+    {
+      "subset_type": "gender",
+      "subset": "",
+      "cer": 0.07329957949855137,
+      "wer": 0.2582569236526946
+    },
+    {
+      "subset_type": "gender",
+      "subset": "other",
+      "cer": 0.0776147657359205,
+      "wer": 0.2745664739884393
+    },
+    {
+      "subset_type": "gender",
+      "subset": "male",
+      "cer": 0.08202917872346872,
+      "wer": 0.28194050672787
+    },
+    {
+      "subset_type": "gender",
+      "subset": "female",
+      "cer": 0.09037798679675177,
+      "wer": 0.30320648343904155
+    },
+    {
+      "subset_type": "accent",
+      "subset": "",
+      "cer": 0.07407837694834264,
+      "wer": 0.2603352040388965
+    },
+    {
+      "subset_type": "accent",
+      "subset": "cameroon",
+      "cer": 0.19595959595959597,
+      "wer": 0.4939759036144578
+    },
+    {
+      "subset_type": "accent",
+      "subset": "canada",
+      "cer": 0.12020629463104998,
+      "wer": 0.3642638036809816
+    },
+    {
+      "subset_type": "accent",
+      "subset": "guadeloupe",
+      "cer": 0.07302709069493522,
+      "wer": 0.3055555555555556
+    },
+    {
+      "subset_type": "accent",
+      "subset": "italy",
+      "cer": 0.18736842105263157,
+      "wer": 0.4936708860759494
+    },
+    {
+      "subset_type": "accent",
+      "subset": "belgium",
+      "cer": 0.09225463428653542,
+      "wer": 0.2848739495798319
+    },
+    {
+      "subset_type": "accent",
+      "subset": "luxembourg",
+      "cer": 0.12368421052631579,
+      "wer": 0.4166666666666667
+    },
+    {
+      "subset_type": "accent",
+      "subset": "romania",
+      "cer": 0.14728682170542637,
+      "wer": 0.5111111111111111
+    },
+    {
+      "subset_type": "accent",
+      "subset": "monaco",
+      "cer": 0.0273224043715847,
+      "wer": 0.15873015873015872
+    },
+    {
+      "subset_type": "accent",
+      "subset": "germany",
+      "cer": 0.06375838926174497,
+      "wer": 0.2830188679245283
+    },
+    {
+      "subset_type": "accent",
+      "subset": "united_states",
+      "cer": 0.11180124223602485,
+      "wer": 0.3292307692307692
+    },
+    {
+      "subset_type": "accent",
+      "subset": "morocco",
+      "cer": 0.05351170568561873,
+      "wer": 0.1346153846153846
+    },
+    {
+      "subset_type": "accent",
+      "subset": "martinique",
+      "cer": 0.0967741935483871,
+      "wer": 0.34210526315789475
+    },
+    {
+      "subset_type": "accent",
+      "subset": "switzerland",
+      "cer": 0.06840390879478828,
+      "wer": 0.22714681440443213
+    },
+    {
+      "subset_type": "accent",
+      "subset": "senegal",
+      "cer": 0.08741258741258741,
+      "wer": 0.32
+    },
+    {
+      "subset_type": "accent",
+      "subset": "gabon",
+      "cer": 0.08888888888888889,
+      "wer": 0.3125
+    },
+    {
+      "subset_type": "accent",
+      "subset": "united_kingdom",
+      "cer": 0.08856088560885608,
+      "wer": 0.2391304347826087
+    },
+    {
+      "subset_type": "accent",
+      "subset": "mayotte",
+      "cer": 0.008583690987124463,
+      "wer": 0.07894736842105263
+    },
+    {
+      "subset_type": "accent",
+      "subset": "france",
+      "cer": 0.07535571266922934,
+      "wer": 0.26800057717281517
+    },
+    {
+      "subset_type": "accent",
+      "subset": "st_pierre_et_miquelon",
+      "cer": 0.156794425087108,
+      "wer": 0.5416666666666666
+    },
+    {
+      "subset_type": "accent",
+      "subset": "cote_d_ivoire",
+      "cer": 0.2256020278833967,
+      "wer": 0.5757575757575758
+    },
+    {
+      "subset_type": "accent",
+      "subset": "tunisia",
+      "cer": 0.19256756756756757,
+      "wer": 0.4375
+    },
+    {
+      "subset_type": "accent",
+      "subset": "algeria",
+      "cer": 0.09585121602288985,
+      "wer": 0.328
+    }
+  ],
+  "test": [
+    {
+      "subset_type": "all",
+      "subset": "all",
+      "cer": 0.08558461820271986,
+      "wer": 0.28212522420316766
+    },
+    {
+      "subset_type": "age",
+      "subset": "",
+      "cer": 0.08444327736200466,
+      "wer": 0.2790159443960245
+    },
+    {
+      "subset_type": "age",
+      "subset": "twenties",
+      "cer": 0.10116454999770758,
+      "wer": 0.3177211722815667
+    },
+    {
+      "subset_type": "age",
+      "subset": "fourties",
+      "cer": 0.08641836390232925,
+      "wer": 0.28879310344827586
+    },
+    {
+      "subset_type": "age",
+      "subset": "thirties",
+      "cer": 0.09479293681976245,
+      "wer": 0.3047982551799346
+    },
+    {
+      "subset_type": "age",
+      "subset": "fifties",
+      "cer": 0.08030746705710103,
+      "wer": 0.2763796909492274
+    },
+    {
+      "subset_type": "age",
+      "subset": "sixties",
+      "cer": 0.08422884819131224,
+      "wer": 0.2816196542311192
+    },
+    {
+      "subset_type": "age",
+      "subset": "seventies",
+      "cer": 0.08862545865086086,
+      "wer": 0.2830508474576271
+    },
+    {
+      "subset_type": "age",
+      "subset": "teens",
+      "cer": 0.09452996562546705,
+      "wer": 0.3160997732426304
+    },
+    {
+      "subset_type": "gender",
+      "subset": "",
+      "cer": 0.08435233361996512,
+      "wer": 0.27890011633947714
+    },
+    {
+      "subset_type": "gender",
+      "subset": "other",
+      "cer": 0.15595880333496812,
+      "wer": 0.4147058823529412
+    },
+    {
+      "subset_type": "gender",
+      "subset": "male",
+      "cer": 0.09267250152656219,
+      "wer": 0.2996407477318395
+    },
+    {
+      "subset_type": "gender",
+      "subset": "female",
+      "cer": 0.09223911304149046,
+      "wer": 0.30696711887790296
+    },
+    {
+      "subset_type": "accent",
+      "subset": "",
+      "cer": 0.08524824440196097,
+      "wer": 0.2809512275052363
+    },
+    {
+      "subset_type": "accent",
+      "subset": "other",
+      "cer": 0.009345794392523364,
+      "wer": 0.1
+    },
+    {
+      "subset_type": "accent",
+      "subset": "cameroon",
+      "cer": 0.13111111111111112,
+      "wer": 0.36363636363636365
+    },
+    {
+      "subset_type": "accent",
+      "subset": "canada",
+      "cer": 0.13479623824451412,
+      "wer": 0.38389830508474576
+    },
+    {
+      "subset_type": "accent",
+      "subset": "reunion",
+      "cer": 0.08150470219435736,
+      "wer": 0.2222222222222222
+    },
+    {
+      "subset_type": "accent",
+      "subset": "guadeloupe",
+      "cer": 0.019762845849802372,
+      "wer": 0.11904761904761904
+    },
+    {
+      "subset_type": "accent",
+      "subset": "benin",
+      "cer": 0.09090909090909091,
+      "wer": 0.2631578947368421
+    },
+    {
+      "subset_type": "accent",
+      "subset": "belgium",
+      "cer": 0.0717928599450765,
+      "wer": 0.25990675990675993
+    },
+    {
+      "subset_type": "accent",
+      "subset": "romania",
+      "cer": 0.0916030534351145,
+      "wer": 0.22727272727272727
+    },
+    {
+      "subset_type": "accent",
+      "subset": "monaco",
+      "cer": 0.09422492401215805,
+      "wer": 0.2641509433962264
+    },
+    {
+      "subset_type": "accent",
+      "subset": "germany",
+      "cer": 0.10175202156334232,
+      "wer": 0.3319838056680162
+    },
+    {
+      "subset_type": "accent",
+      "subset": "united_states",
+      "cer": 0.21368948247078465,
+      "wer": 0.5737179487179487
+    },
+    {
+      "subset_type": "accent",
+      "subset": "morocco",
+      "cer": 0.0784313725490196,
+      "wer": 0.27586206896551724
+    },
+    {
+      "subset_type": "accent",
+      "subset": "martinique",
+      "cer": 0.03070175438596491,
+      "wer": 0.10526315789473684
+    },
+    {
+      "subset_type": "accent",
+      "subset": "burundi",
+      "cer": 0.10714285714285714,
+      "wer": 0.35714285714285715
+    },
+    {
+      "subset_type": "accent",
+      "subset": "switzerland",
+      "cer": 0.07480073574494175,
+      "wer": 0.26277372262773724
+    },
+    {
+      "subset_type": "accent",
+      "subset": "senegal",
+      "cer": 0.11782477341389729,
+      "wer": 0.34513274336283184
+    },
+    {
+      "subset_type": "accent",
+      "subset": "united_kingdom",
+      "cer": 0.17142857142857143,
+      "wer": 0.3333333333333333
+    },
+    {
+      "subset_type": "accent",
+      "subset": "france",
+      "cer": 0.08173514857578966,
+      "wer": 0.277514090774251
+    },
+    {
+      "subset_type": "accent",
+      "subset": "cote_d_ivoire",
+      "cer": 0.1252847380410023,
+      "wer": 0.38461538461538464
+    },
+    {
+      "subset_type": "accent",
+      "subset": "new_caledonia",
+      "cer": 0.0959409594095941,
+      "wer": 0.2926829268292683
+    },
+    {
+      "subset_type": "accent",
+      "subset": "tunisia",
+      "cer": 0.10458715596330276,
+      "wer": 0.3191489361702128
+    },
+    {
+      "subset_type": "accent",
+      "subset": "algeria",
+      "cer": 0.11220825852782765,
+      "wer": 0.3817204301075269
+    }
+  ]
+}

cv7_transcriptions.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f66d5eb3eb7aac9494947298c139c396c988a39292123982da9966daae06af3b
+size 130413755

preprocessor_config.json

@@ -0,0 +1,10 @@
+{
+  "do_normalize": true,
+  "feature_extractor_type": "Wav2Vec2FeatureExtractor",
+  "feature_size": 1,
+  "padding_side": "right",
+  "padding_value": 0,
+  "processor_class": "Wav2Vec2Processor",
+  "return_attention_mask": true,
+  "sampling_rate": 16000
+}

pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:42f414d38a5abcaa91398d803b563c897f43dee74b0281f8eceb7d83dd0d9930
+size 1262132401

special_tokens_map.json

@@ -0,0 +1 @@
+{"bos_token": "<s>", "eos_token": "</s>", "unk_token": "<unk>", "pad_token": "<pad>"}

tokenizer_config.json

@@ -0,0 +1 @@
+{"unk_token": "<unk>", "bos_token": "<s>", "eos_token": "</s>", "pad_token": "<pad>", "do_lower_case": false, "word_delimiter_token": "|", "replace_word_delimiter_char": " ", "special_tokens_map_file": null, "name_or_path": "/tmp/tmp8rg3fryf", "tokenizer_class": "Wav2Vec2CTCTokenizer", "processor_class": "Wav2Vec2Processor"}

vocab.json

@@ -0,0 +1 @@
+{"<pad>": 0, "<s>": 1, "</s>": 2, "<unk>": 3, "|": 4, "'": 5, "-": 6, "a": 7, "b": 8, "c": 9, "d": 10, "e": 11, "f": 12, "g": 13, "h": 14, "i": 15, "j": 16, "k": 17, "l": 18, "m": 19, "n": 20, "o": 21, "p": 22, "q": 23, "r": 24, "s": 25, "t": 26, "u": 27, "v": 28, "w": 29, "x": 30, "y": 31, "z": 32, "à": 33, "á": 34, "â": 35, "ä": 36, "ç": 37, "è": 38, "é": 39, "ê": 40, "ë": 41, "í": 42, "î": 43, "ï": 44, "ó": 45, "ô": 46, "ö": 47, "ù": 48, "û": 49, "ü": 50, "ć": 51, "č": 52, "ł": 53, "ō": 54, "œ": 55, "š": 56}