jonatasgrosman/exp_w2v2r_fr_vp-100k_age_teens-10_sixties-0_s451

.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_s451
+
+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": 0.10487911789335375,
+      "wer": 0.3485325784829296
+    },
+    {
+      "subset_type": "age",
+      "subset": "",
+      "cer": 0.10255279926261371,
+      "wer": 0.34319766625586945
+    },
+    {
+      "subset_type": "age",
+      "subset": "thirties",
+      "cer": 0.11367222868969086,
+      "wer": 0.3662969808995687
+    },
+    {
+      "subset_type": "age",
+      "subset": "sixties",
+      "cer": 0.10238311552169962,
+      "wer": 0.3347602739726027
+    },
+    {
+      "subset_type": "age",
+      "subset": "seventies",
+      "cer": 0.11295734142834318,
+      "wer": 0.3668582375478927
+    },
+    {
+      "subset_type": "age",
+      "subset": "fifties",
+      "cer": 0.12027890761185357,
+      "wer": 0.3684210526315789
+    },
+    {
+      "subset_type": "age",
+      "subset": "twenties",
+      "cer": 0.11804857019686736,
+      "wer": 0.38414699284264503
+    },
+    {
+      "subset_type": "age",
+      "subset": "eighties",
+      "cer": 0.09142857142857143,
+      "wer": 0.41379310344827586
+    },
+    {
+      "subset_type": "age",
+      "subset": "teens",
+      "cer": 0.1159511519576986,
+      "wer": 0.3724770642201835
+    },
+    {
+      "subset_type": "age",
+      "subset": "fourties",
+      "cer": 0.11862160371106693,
+      "wer": 0.3813500347947112
+    },
+    {
+      "subset_type": "gender",
+      "subset": "",
+      "cer": 0.10255376448300549,
+      "wer": 0.3431808258483034
+    },
+    {
+      "subset_type": "gender",
+      "subset": "male",
+      "cer": 0.11431363340251792,
+      "wer": 0.37046037532064263
+    },
+    {
+      "subset_type": "gender",
+      "subset": "other",
+      "cer": 0.09891150023663038,
+      "wer": 0.3236994219653179
+    },
+    {
+      "subset_type": "gender",
+      "subset": "female",
+      "cer": 0.1214289887246597,
+      "wer": 0.38513037350246654
+    },
+    {
+      "subset_type": "accent",
+      "subset": "",
+      "cer": 0.10359759108261112,
+      "wer": 0.34557541329393937
+    },
+    {
+      "subset_type": "accent",
+      "subset": "united_kingdom",
+      "cer": 0.19557195571955718,
+      "wer": 0.7391304347826086
+    },
+    {
+      "subset_type": "accent",
+      "subset": "st_pierre_et_miquelon",
+      "cer": 0.24390243902439024,
+      "wer": 0.7916666666666666
+    },
+    {
+      "subset_type": "accent",
+      "subset": "martinique",
+      "cer": 0.18548387096774194,
+      "wer": 0.42105263157894735
+    },
+    {
+      "subset_type": "accent",
+      "subset": "romania",
+      "cer": 0.14728682170542637,
+      "wer": 0.5333333333333333
+    },
+    {
+      "subset_type": "accent",
+      "subset": "cote_d_ivoire",
+      "cer": 0.3231939163498099,
+      "wer": 0.7045454545454546
+    },
+    {
+      "subset_type": "accent",
+      "subset": "guadeloupe",
+      "cer": 0.0989399293286219,
+      "wer": 0.3263888888888889
+    },
+    {
+      "subset_type": "accent",
+      "subset": "switzerland",
+      "cer": 0.08748255002326663,
+      "wer": 0.2853185595567867
+    },
+    {
+      "subset_type": "accent",
+      "subset": "united_states",
+      "cer": 0.12267080745341614,
+      "wer": 0.36615384615384616
+    },
+    {
+      "subset_type": "accent",
+      "subset": "canada",
+      "cer": 0.15577889447236182,
+      "wer": 0.4532208588957055
+    },
+    {
+      "subset_type": "accent",
+      "subset": "morocco",
+      "cer": 0.06020066889632107,
+      "wer": 0.23076923076923078
+    },
+    {
+      "subset_type": "accent",
+      "subset": "belgium",
+      "cer": 0.12444316712171288,
+      "wer": 0.373109243697479
+    },
+    {
+      "subset_type": "accent",
+      "subset": "luxembourg",
+      "cer": 0.12631578947368421,
+      "wer": 0.4
+    },
+    {
+      "subset_type": "accent",
+      "subset": "italy",
+      "cer": 0.2694736842105263,
+      "wer": 0.5569620253164557
+    },
+    {
+      "subset_type": "accent",
+      "subset": "monaco",
+      "cer": 0.060109289617486336,
+      "wer": 0.31746031746031744
+    },
+    {
+      "subset_type": "accent",
+      "subset": "senegal",
+      "cer": 0.17132867132867133,
+      "wer": 0.48
+    },
+    {
+      "subset_type": "accent",
+      "subset": "mayotte",
+      "cer": 0.04721030042918455,
+      "wer": 0.21052631578947367
+    },
+    {
+      "subset_type": "accent",
+      "subset": "france",
+      "cer": 0.10513686812349028,
+      "wer": 0.35246019912462123
+    },
+    {
+      "subset_type": "accent",
+      "subset": "germany",
+      "cer": 0.09899328859060402,
+      "wer": 0.37735849056603776
+    },
+    {
+      "subset_type": "accent",
+      "subset": "tunisia",
+      "cer": 0.22635135135135134,
+      "wer": 0.4791666666666667
+    },
+    {
+      "subset_type": "accent",
+      "subset": "gabon",
+      "cer": 0.1,
+      "wer": 0.3333333333333333
+    },
+    {
+      "subset_type": "accent",
+      "subset": "algeria",
+      "cer": 0.1630901287553648,
+      "wer": 0.528
+    },
+    {
+      "subset_type": "accent",
+      "subset": "cameroon",
+      "cer": 0.24242424242424243,
+      "wer": 0.5903614457831325
+    }
+  ],
+  "test": [
+    {
+      "subset_type": "all",
+      "subset": "all",
+      "cer": 0.11765845318780253,
+      "wer": 0.3677629118841939
+    },
+    {
+      "subset_type": "age",
+      "subset": "",
+      "cer": 0.11639215915114204,
+      "wer": 0.36391600811501645
+    },
+    {
+      "subset_type": "age",
+      "subset": "thirties",
+      "cer": 0.1290701350313869,
+      "wer": 0.3964013086150491
+    },
+    {
+      "subset_type": "age",
+      "subset": "seventies",
+      "cer": 0.11882585379621789,
+      "wer": 0.3813559322033898
+    },
+    {
+      "subset_type": "age",
+      "subset": "fifties",
+      "cer": 0.11076134699853588,
+      "wer": 0.3637969094922737
+    },
+    {
+      "subset_type": "age",
+      "subset": "twenties",
+      "cer": 0.12977396726422447,
+      "wer": 0.4044097507532183
+    },
+    {
+      "subset_type": "age",
+      "subset": "fourties",
+      "cer": 0.1263643524248903,
+      "wer": 0.3952254641909814
+    },
+    {
+      "subset_type": "age",
+      "subset": "teens",
+      "cer": 0.12561649977581826,
+      "wer": 0.3832199546485261
+    },
+    {
+      "subset_type": "age",
+      "subset": "sixties",
+      "cer": 0.12441350083245044,
+      "wer": 0.38444040036396726
+    },
+    {
+      "subset_type": "gender",
+      "subset": "",
+      "cer": 0.11629720315392127,
+      "wer": 0.3637840329555349
+    },
+    {
+      "subset_type": "gender",
+      "subset": "male",
+      "cer": 0.12521880724608184,
+      "wer": 0.38969737563173595
+    },
+    {
+      "subset_type": "gender",
+      "subset": "other",
+      "cer": 0.19077979401667483,
+      "wer": 0.4647058823529412
+    },
+    {
+      "subset_type": "gender",
+      "subset": "female",
+      "cer": 0.12637634777307882,
+      "wer": 0.4023913543343297
+    },
+    {
+      "subset_type": "accent",
+      "subset": "",
+      "cer": 0.1173340462985747,
+      "wer": 0.36639927047860593
+    },
+    {
+      "subset_type": "accent",
+      "subset": "united_kingdom",
+      "cer": 0.34285714285714286,
+      "wer": 0.6666666666666666
+    },
+    {
+      "subset_type": "accent",
+      "subset": "benin",
+      "cer": 0.10454545454545454,
+      "wer": 0.39473684210526316
+    },
+    {
+      "subset_type": "accent",
+      "subset": "reunion",
+      "cer": 0.09717868338557993,
+      "wer": 0.2777777777777778
+    },
+    {
+      "subset_type": "accent",
+      "subset": "martinique",
+      "cer": 0.07894736842105263,
+      "wer": 0.2894736842105263
+    },
+    {
+      "subset_type": "accent",
+      "subset": "romania",
+      "cer": 0.183206106870229,
+      "wer": 0.5454545454545454
+    },
+    {
+      "subset_type": "accent",
+      "subset": "cote_d_ivoire",
+      "cer": 0.20045558086560364,
+      "wer": 0.5314685314685315
+    },
+    {
+      "subset_type": "accent",
+      "subset": "other",
+      "cer": 0.04672897196261682,
+      "wer": 0.3
+    },
+    {
+      "subset_type": "accent",
+      "subset": "guadeloupe",
+      "cer": 0.07509881422924901,
+      "wer": 0.23809523809523808
+    },
+    {
+      "subset_type": "accent",
+      "subset": "switzerland",
+      "cer": 0.10300429184549356,
+      "wer": 0.35036496350364965
+    },
+    {
+      "subset_type": "accent",
+      "subset": "united_states",
+      "cer": 0.25208681135225375,
+      "wer": 0.6153846153846154
+    },
+    {
+      "subset_type": "accent",
+      "subset": "canada",
+      "cer": 0.17269877457965233,
+      "wer": 0.48898305084745763
+    },
+    {
+      "subset_type": "accent",
+      "subset": "burundi",
+      "cer": 0.19047619047619047,
+      "wer": 0.6428571428571429
+    },
+    {
+      "subset_type": "accent",
+      "subset": "morocco",
+      "cer": 0.08823529411764706,
+      "wer": 0.3017241379310345
+    },
+    {
+      "subset_type": "accent",
+      "subset": "belgium",
+      "cer": 0.11377010592389172,
+      "wer": 0.38694638694638694
+    },
+    {
+      "subset_type": "accent",
+      "subset": "monaco",
+      "cer": 0.11550151975683891,
+      "wer": 0.3584905660377358
+    },
+    {
+      "subset_type": "accent",
+      "subset": "senegal",
+      "cer": 0.1148036253776435,
+      "wer": 0.35398230088495575
+    },
+    {
+      "subset_type": "accent",
+      "subset": "france",
+      "cer": 0.11286590019449026,
+      "wer": 0.3611687926431326
+    },
+    {
+      "subset_type": "accent",
+      "subset": "germany",
+      "cer": 0.11522911051212938,
+      "wer": 0.41295546558704455
+    },
+    {
+      "subset_type": "accent",
+      "subset": "tunisia",
+      "cer": 0.13211009174311927,
+      "wer": 0.43617021276595747
+    },
+    {
+      "subset_type": "accent",
+      "subset": "algeria",
+      "cer": 0.11759425493716337,
+      "wer": 0.42473118279569894
+    },
+    {
+      "subset_type": "accent",
+      "subset": "cameroon",
+      "cer": 0.1622222222222222,
+      "wer": 0.4675324675324675
+    },
+    {
+      "subset_type": "accent",
+      "subset": "new_caledonia",
+      "cer": 0.1918819188191882,
+      "wer": 0.4146341463414634
+    }
+  ]
+}

cv7_transcriptions.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bbd3d02179d1e67d259fde636d7c13b205e456e2d37cde5ce1f1174eb520297d
+size 128960013

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:bae03d7fbc5b62bb465d406e05a31aa5f1a7614dec45f1f9867785f1ca62ded3
+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/tmpxczis8_h", "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}