jonatasgrosman/exp_w2v2r_fr_vp-100k_age_teens-10_sixties-0_s818

.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_vp-100k_age_teens-10_sixties-0_s818
+
+Fine-tuned [facebook/wav2vec2-large-100k-voxpopuli](https://huggingface.co/facebook/wav2vec2-large-100k-voxpopuli) 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 +1,116 @@
+{
+  "_name_or_path": "facebook/wav2vec2-large-100k-voxpopuli",
+  "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_channel_length": 10,
+  "mask_channel_min_space": 1,
+  "mask_channel_other": 0.0,
+  "mask_channel_prob": 0.0,
+  "mask_channel_selection": "static",
+  "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_min_space": 1,
+  "mask_time_other": 0.0,
+  "mask_time_prob": 0.05,
+  "mask_time_selection": "static",
+  "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": 0.10440894121093315,
+      "wer": 0.3464431082626726
+    },
+    {
+      "subset_type": "age",
+      "subset": "",
+      "cer": 0.10223478203538781,
+      "wer": 0.34119307988705677
+    },
+    {
+      "subset_type": "age",
+      "subset": "thirties",
+      "cer": 0.11354287931703531,
+      "wer": 0.3679913739987677
+    },
+    {
+      "subset_type": "age",
+      "subset": "sixties",
+      "cer": 0.10175414075057657,
+      "wer": 0.336472602739726
+    },
+    {
+      "subset_type": "age",
+      "subset": "seventies",
+      "cer": 0.11231826170314747,
+      "wer": 0.36590038314176243
+    },
+    {
+      "subset_type": "age",
+      "subset": "fifties",
+      "cer": 0.12169004731468415,
+      "wer": 0.36541353383458647
+    },
+    {
+      "subset_type": "age",
+      "subset": "twenties",
+      "cer": 0.11677324328208076,
+      "wer": 0.3807285546415981
+    },
+    {
+      "subset_type": "age",
+      "subset": "eighties",
+      "cer": 0.09714285714285714,
+      "wer": 0.3793103448275862
+    },
+    {
+      "subset_type": "age",
+      "subset": "teens",
+      "cer": 0.11053758025934786,
+      "wer": 0.3541284403669725
+    },
+    {
+      "subset_type": "age",
+      "subset": "fourties",
+      "cer": 0.1170623318910069,
+      "wer": 0.38250985850150776
+    },
+    {
+      "subset_type": "gender",
+      "subset": "",
+      "cer": 0.10223974526968177,
+      "wer": 0.3411926147704591
+    },
+    {
+      "subset_type": "gender",
+      "subset": "male",
+      "cer": 0.11228354568454861,
+      "wer": 0.3665451599837991
+    },
+    {
+      "subset_type": "gender",
+      "subset": "other",
+      "cer": 0.10317084713677237,
+      "wer": 0.34104046242774566
+    },
+    {
+      "subset_type": "gender",
+      "subset": "female",
+      "cer": 0.12315242156920021,
+      "wer": 0.38671599718111344
+    },
+    {
+      "subset_type": "accent",
+      "subset": "",
+      "cer": 0.10319980906463137,
+      "wer": 0.34347372509004515
+    },
+    {
+      "subset_type": "accent",
+      "subset": "united_kingdom",
+      "cer": 0.2066420664206642,
+      "wer": 0.7391304347826086
+    },
+    {
+      "subset_type": "accent",
+      "subset": "st_pierre_et_miquelon",
+      "cer": 0.30662020905923343,
+      "wer": 0.75
+    },
+    {
+      "subset_type": "accent",
+      "subset": "martinique",
+      "cer": 0.14516129032258066,
+      "wer": 0.39473684210526316
+    },
+    {
+      "subset_type": "accent",
+      "subset": "romania",
+      "cer": 0.15503875968992248,
+      "wer": 0.5555555555555556
+    },
+    {
+      "subset_type": "accent",
+      "subset": "cote_d_ivoire",
+      "cer": 0.3193916349809886,
+      "wer": 0.6893939393939394
+    },
+    {
+      "subset_type": "accent",
+      "subset": "guadeloupe",
+      "cer": 0.0989399293286219,
+      "wer": 0.3263888888888889
+    },
+    {
+      "subset_type": "accent",
+      "subset": "switzerland",
+      "cer": 0.10190786412284783,
+      "wer": 0.31301939058171746
+    },
+    {
+      "subset_type": "accent",
+      "subset": "united_states",
+      "cer": 0.12888198757763975,
+      "wer": 0.3630769230769231
+    },
+    {
+      "subset_type": "accent",
+      "subset": "canada",
+      "cer": 0.15670457550912456,
+      "wer": 0.46855828220858897
+    },
+    {
+      "subset_type": "accent",
+      "subset": "morocco",
+      "cer": 0.056856187290969896,
+      "wer": 0.25
+    },
+    {
+      "subset_type": "accent",
+      "subset": "belgium",
+      "cer": 0.1222876850122144,
+      "wer": 0.3739495798319328
+    },
+    {
+      "subset_type": "accent",
+      "subset": "luxembourg",
+      "cer": 0.1394736842105263,
+      "wer": 0.38333333333333336
+    },
+    {
+      "subset_type": "accent",
+      "subset": "italy",
+      "cer": 0.23578947368421052,
+      "wer": 0.5189873417721519
+    },
+    {
+      "subset_type": "accent",
+      "subset": "monaco",
+      "cer": 0.06284153005464481,
+      "wer": 0.31746031746031744
+    },
+    {
+      "subset_type": "accent",
+      "subset": "senegal",
+      "cer": 0.15384615384615385,
+      "wer": 0.46
+    },
+    {
+      "subset_type": "accent",
+      "subset": "mayotte",
+      "cer": 0.034334763948497854,
+      "wer": 0.18421052631578946
+    },
+    {
+      "subset_type": "accent",
+      "subset": "france",
+      "cer": 0.10377260069497388,
+      "wer": 0.3489971622336588
+    },
+    {
+      "subset_type": "accent",
+      "subset": "germany",
+      "cer": 0.10234899328859061,
+      "wer": 0.4056603773584906
+    },
+    {
+      "subset_type": "accent",
+      "subset": "tunisia",
+      "cer": 0.24324324324324326,
+      "wer": 0.5208333333333334
+    },
+    {
+      "subset_type": "accent",
+      "subset": "gabon",
+      "cer": 0.1,
+      "wer": 0.3541666666666667
+    },
+    {
+      "subset_type": "accent",
+      "subset": "algeria",
+      "cer": 0.1530758226037196,
+      "wer": 0.472
+    },
+    {
+      "subset_type": "accent",
+      "subset": "cameroon",
+      "cer": 0.27474747474747474,
+      "wer": 0.6024096385542169
+    }
+  ],
+  "test": [
+    {
+      "subset_type": "all",
+      "subset": "all",
+      "cer": 0.11820482533987717,
+      "wer": 0.36746503064374036
+    },
+    {
+      "subset_type": "age",
+      "subset": "",
+      "cer": 0.11694652331004031,
+      "wer": 0.36351318690171897
+    },
+    {
+      "subset_type": "age",
+      "subset": "thirties",
+      "cer": 0.12979271101476764,
+      "wer": 0.40239912758996726
+    },
+    {
+      "subset_type": "age",
+      "subset": "seventies",
+      "cer": 0.11543889359300029,
+      "wer": 0.3728813559322034
+    },
+    {
+      "subset_type": "age",
+      "subset": "fifties",
+      "cer": 0.11434846266471449,
+      "wer": 0.36423841059602646
+    },
+    {
+      "subset_type": "age",
+      "subset": "twenties",
+      "cer": 0.1293613314382651,
+      "wer": 0.40221857025472474
+    },
+    {
+      "subset_type": "age",
+      "subset": "fourties",
+      "cer": 0.12664566220321818,
+      "wer": 0.3905835543766578
+    },
+    {
+      "subset_type": "age",
+      "subset": "teens",
+      "cer": 0.126513226722463,
+      "wer": 0.3963718820861678
+    },
+    {
+      "subset_type": "age",
+      "subset": "sixties",
+      "cer": 0.1252459512638111,
+      "wer": 0.38034576888080074
+    },
+    {
+      "subset_type": "gender",
+      "subset": "",
+      "cer": 0.11684010318744262,
+      "wer": 0.3634035516466792
+    },
+    {
+      "subset_type": "gender",
+      "subset": "male",
+      "cer": 0.1251984530836556,
+      "wer": 0.3895147049869086
+    },
+    {
+      "subset_type": "gender",
+      "subset": "other",
+      "cer": 0.19813634134379599,
+      "wer": 0.4823529411764706
+    },
+    {
+      "subset_type": "gender",
+      "subset": "female",
+      "cer": 0.12862422372080617,
+      "wer": 0.4028512301678547
+    },
+    {
+      "subset_type": "accent",
+      "subset": "",
+      "cer": 0.11789716265071644,
+      "wer": 0.36615704657823117
+    },
+    {
+      "subset_type": "accent",
+      "subset": "united_kingdom",
+      "cer": 0.2857142857142857,
+      "wer": 0.6666666666666666
+    },
+    {
+      "subset_type": "accent",
+      "subset": "benin",
+      "cer": 0.1,
+      "wer": 0.3684210526315789
+    },
+    {
+      "subset_type": "accent",
+      "subset": "reunion",
+      "cer": 0.10031347962382445,
+      "wer": 0.3333333333333333
+    },
+    {
+      "subset_type": "accent",
+      "subset": "martinique",
+      "cer": 0.10526315789473684,
+      "wer": 0.2894736842105263
+    },
+    {
+      "subset_type": "accent",
+      "subset": "romania",
+      "cer": 0.16030534351145037,
+      "wer": 0.4090909090909091
+    },
+    {
+      "subset_type": "accent",
+      "subset": "cote_d_ivoire",
+      "cer": 0.2266514806378132,
+      "wer": 0.5874125874125874
+    },
+    {
+      "subset_type": "accent",
+      "subset": "other",
+      "cer": 0.04672897196261682,
+      "wer": 0.35
+    },
+    {
+      "subset_type": "accent",
+      "subset": "guadeloupe",
+      "cer": 0.05138339920948617,
+      "wer": 0.23809523809523808
+    },
+    {
+      "subset_type": "accent",
+      "subset": "switzerland",
+      "cer": 0.11158798283261803,
+      "wer": 0.35766423357664234
+    },
+    {
+      "subset_type": "accent",
+      "subset": "united_states",
+      "cer": 0.25041736227045075,
+      "wer": 0.6057692307692307
+    },
+    {
+      "subset_type": "accent",
+      "subset": "canada",
+      "cer": 0.17155884867483615,
+      "wer": 0.47627118644067795
+    },
+    {
+      "subset_type": "accent",
+      "subset": "burundi",
+      "cer": 0.15476190476190477,
+      "wer": 0.5714285714285714
+    },
+    {
+      "subset_type": "accent",
+      "subset": "morocco",
+      "cer": 0.08823529411764706,
+      "wer": 0.31896551724137934
+    },
+    {
+      "subset_type": "accent",
+      "subset": "belgium",
+      "cer": 0.11239701843860338,
+      "wer": 0.36596736596736595
+    },
+    {
+      "subset_type": "accent",
+      "subset": "monaco",
+      "cer": 0.11246200607902736,
+      "wer": 0.39622641509433965
+    },
+    {
+      "subset_type": "accent",
+      "subset": "senegal",
+      "cer": 0.1283987915407855,
+      "wer": 0.336283185840708
+    },
+    {
+      "subset_type": "accent",
+      "subset": "france",
+      "cer": 0.11338289962825279,
+      "wer": 0.36287451794719666
+    },
+    {
+      "subset_type": "accent",
+      "subset": "germany",
+      "cer": 0.11590296495956873,
+      "wer": 0.4008097165991903
+    },
+    {
+      "subset_type": "accent",
+      "subset": "tunisia",
+      "cer": 0.11376146788990826,
+      "wer": 0.4148936170212766
+    },
+    {
+      "subset_type": "accent",
+      "subset": "algeria",
+      "cer": 0.09784560143626571,
+      "wer": 0.3709677419354839
+    },
+    {
+      "subset_type": "accent",
+      "subset": "cameroon",
+      "cer": 0.16444444444444445,
+      "wer": 0.4805194805194805
+    },
+    {
+      "subset_type": "accent",
+      "subset": "new_caledonia",
+      "cer": 0.23616236162361623,
+      "wer": 0.4878048780487805
+    }
+  ]
+}

cv7_transcriptions.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2e30c252e9deb2550240405f5c97d0ef390557c19a647c96e260876550c339e2
+size 129406214

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 sha256:6749810020d873152eca3f7c3210699a192721195121f389e190f9e4261a5c45
+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/tmpsyuwde5d", "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}