jonatasgrosman/exp_w2v2r_fr_vp-100k_age_teens-5_sixties-5_s693

.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-5_sixties-5_s693
+
+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": 0.10319369899564741,
+      "wer": 0.34185521952216946
+    },
+    {
+      "subset_type": "age",
+      "subset": "",
+      "cer": 0.10082697406608294,
+      "wer": 0.33682510958925477
+    },
+    {
+      "subset_type": "age",
+      "subset": "thirties",
+      "cer": 0.11206829646876212,
+      "wer": 0.3556685150955022
+    },
+    {
+      "subset_type": "age",
+      "subset": "sixties",
+      "cer": 0.09574393738206723,
+      "wer": 0.3227739726027397
+    },
+    {
+      "subset_type": "age",
+      "subset": "seventies",
+      "cer": 0.10832401342067423,
+      "wer": 0.35919540229885055
+    },
+    {
+      "subset_type": "age",
+      "subset": "fifties",
+      "cer": 0.11787166929526023,
+      "wer": 0.35388471177944864
+    },
+    {
+      "subset_type": "age",
+      "subset": "twenties",
+      "cer": 0.11858744072424199,
+      "wer": 0.37549407114624506
+    },
+    {
+      "subset_type": "age",
+      "subset": "eighties",
+      "cer": 0.09714285714285714,
+      "wer": 0.3103448275862069
+    },
+    {
+      "subset_type": "age",
+      "subset": "teens",
+      "cer": 0.11853204079063326,
+      "wer": 0.3772477064220183
+    },
+    {
+      "subset_type": "age",
+      "subset": "fourties",
+      "cer": 0.11433360620590184,
+      "wer": 0.376014845743447
+    },
+    {
+      "subset_type": "gender",
+      "subset": "",
+      "cer": 0.10082988946006786,
+      "wer": 0.33682634730538924
+    },
+    {
+      "subset_type": "gender",
+      "subset": "male",
+      "cer": 0.11228354568454861,
+      "wer": 0.3616848926690968
+    },
+    {
+      "subset_type": "gender",
+      "subset": "other",
+      "cer": 0.10175106483672504,
+      "wer": 0.3179190751445087
+    },
+    {
+      "subset_type": "gender",
+      "subset": "female",
+      "cer": 0.12166267453408892,
+      "wer": 0.37931642001409444
+    },
+    {
+      "subset_type": "accent",
+      "subset": "",
+      "cer": 0.1017905242009317,
+      "wer": 0.3390875931862659
+    },
+    {
+      "subset_type": "accent",
+      "subset": "united_kingdom",
+      "cer": 0.16236162361623616,
+      "wer": 0.5217391304347826
+    },
+    {
+      "subset_type": "accent",
+      "subset": "st_pierre_et_miquelon",
+      "cer": 0.27177700348432055,
+      "wer": 0.7291666666666666
+    },
+    {
+      "subset_type": "accent",
+      "subset": "martinique",
+      "cer": 0.17338709677419356,
+      "wer": 0.39473684210526316
+    },
+    {
+      "subset_type": "accent",
+      "subset": "romania",
+      "cer": 0.16279069767441862,
+      "wer": 0.6222222222222222
+    },
+    {
+      "subset_type": "accent",
+      "subset": "cote_d_ivoire",
+      "cer": 0.30798479087452474,
+      "wer": 0.696969696969697
+    },
+    {
+      "subset_type": "accent",
+      "subset": "guadeloupe",
+      "cer": 0.10129564193168433,
+      "wer": 0.3472222222222222
+    },
+    {
+      "subset_type": "accent",
+      "subset": "switzerland",
+      "cer": 0.08980921358771522,
+      "wer": 0.27146814404432135
+    },
+    {
+      "subset_type": "accent",
+      "subset": "united_states",
+      "cer": 0.13612836438923395,
+      "wer": 0.37846153846153846
+    },
+    {
+      "subset_type": "accent",
+      "subset": "canada",
+      "cer": 0.16106850039672044,
+      "wer": 0.46702453987730064
+    },
+    {
+      "subset_type": "accent",
+      "subset": "morocco",
+      "cer": 0.05016722408026756,
+      "wer": 0.21153846153846154
+    },
+    {
+      "subset_type": "accent",
+      "subset": "belgium",
+      "cer": 0.1218565885903147,
+      "wer": 0.3722689075630252
+    },
+    {
+      "subset_type": "accent",
+      "subset": "luxembourg",
+      "cer": 0.13421052631578947,
+      "wer": 0.4166666666666667
+    },
+    {
+      "subset_type": "accent",
+      "subset": "italy",
+      "cer": 0.24210526315789474,
+      "wer": 0.569620253164557
+    },
+    {
+      "subset_type": "accent",
+      "subset": "monaco",
+      "cer": 0.07377049180327869,
+      "wer": 0.3333333333333333
+    },
+    {
+      "subset_type": "accent",
+      "subset": "senegal",
+      "cer": 0.17132867132867133,
+      "wer": 0.48
+    },
+    {
+      "subset_type": "accent",
+      "subset": "mayotte",
+      "cer": 0.02145922746781116,
+      "wer": 0.13157894736842105
+    },
+    {
+      "subset_type": "accent",
+      "subset": "france",
+      "cer": 0.1034997472092706,
+      "wer": 0.3427925544706844
+    },
+    {
+      "subset_type": "accent",
+      "subset": "germany",
+      "cer": 0.11241610738255034,
+      "wer": 0.3867924528301887
+    },
+    {
+      "subset_type": "accent",
+      "subset": "tunisia",
+      "cer": 0.2533783783783784,
+      "wer": 0.5416666666666666
+    },
+    {
+      "subset_type": "accent",
+      "subset": "gabon",
+      "cer": 0.10740740740740741,
+      "wer": 0.3333333333333333
+    },
+    {
+      "subset_type": "accent",
+      "subset": "algeria",
+      "cer": 0.1630901287553648,
+      "wer": 0.52
+    },
+    {
+      "subset_type": "accent",
+      "subset": "cameroon",
+      "cer": 0.24242424242424243,
+      "wer": 0.5662650602409639
+    }
+  ],
+  "test": [
+    {
+      "subset_type": "all",
+      "subset": "all",
+      "cer": 0.11633621152497836,
+      "wer": 0.362141195707975
+    },
+    {
+      "subset_type": "age",
+      "subset": "",
+      "cer": 0.11510107420212895,
+      "wer": 0.35859144407742954
+    },
+    {
+      "subset_type": "age",
+      "subset": "thirties",
+      "cer": 0.12861852504177393,
+      "wer": 0.39176663031624864
+    },
+    {
+      "subset_type": "age",
+      "subset": "seventies",
+      "cer": 0.1013265594129269,
+      "wer": 0.3474576271186441
+    },
+    {
+      "subset_type": "age",
+      "subset": "fifties",
+      "cer": 0.10812591508052709,
+      "wer": 0.34392935982339956
+    },
+    {
+      "subset_type": "age",
+      "subset": "twenties",
+      "cer": 0.13018660309018384,
+      "wer": 0.4004382360996987
+    },
+    {
+      "subset_type": "age",
+      "subset": "fourties",
+      "cer": 0.12388882637560482,
+      "wer": 0.3796419098143236
+    },
+    {
+      "subset_type": "age",
+      "subset": "teens",
+      "cer": 0.12718577193244657,
+      "wer": 0.3968253968253968
+    },
+    {
+      "subset_type": "age",
+      "subset": "sixties",
+      "cer": 0.11828363856515817,
+      "wer": 0.36988171064604186
+    },
+    {
+      "subset_type": "gender",
+      "subset": "",
+      "cer": 0.11500736012125982,
+      "wer": 0.3583768082008356
+    },
+    {
+      "subset_type": "gender",
+      "subset": "male",
+      "cer": 0.12345817219621412,
+      "wer": 0.3828776715581806
+    },
+    {
+      "subset_type": "gender",
+      "subset": "other",
+      "cer": 0.19323197645904855,
+      "wer": 0.5058823529411764
+    },
+    {
+      "subset_type": "gender",
+      "subset": "female",
+      "cer": 0.12538575837238541,
+      "wer": 0.390894458496206
+    },
+    {
+      "subset_type": "accent",
+      "subset": "",
+      "cer": 0.11596884404989495,
+      "wer": 0.36093498425544646
+    },
+    {
+      "subset_type": "accent",
+      "subset": "united_kingdom",
+      "cer": 0.34285714285714286,
+      "wer": 0.7777777777777778
+    },
+    {
+      "subset_type": "accent",
+      "subset": "benin",
+      "cer": 0.1,
+      "wer": 0.34210526315789475
+    },
+    {
+      "subset_type": "accent",
+      "subset": "reunion",
+      "cer": 0.12225705329153605,
+      "wer": 0.4074074074074074
+    },
+    {
+      "subset_type": "accent",
+      "subset": "martinique",
+      "cer": 0.10087719298245613,
+      "wer": 0.2894736842105263
+    },
+    {
+      "subset_type": "accent",
+      "subset": "romania",
+      "cer": 0.07633587786259542,
+      "wer": 0.22727272727272727
+    },
+    {
+      "subset_type": "accent",
+      "subset": "cote_d_ivoire",
+      "cer": 0.2072892938496583,
+      "wer": 0.5454545454545454
+    },
+    {
+      "subset_type": "accent",
+      "subset": "other",
+      "cer": 0.07476635514018691,
+      "wer": 0.35
+    },
+    {
+      "subset_type": "accent",
+      "subset": "guadeloupe",
+      "cer": 0.03557312252964427,
+      "wer": 0.16666666666666666
+    },
+    {
+      "subset_type": "accent",
+      "subset": "switzerland",
+      "cer": 0.11955855303494789,
+      "wer": 0.38686131386861317
+    },
+    {
+      "subset_type": "accent",
+      "subset": "united_states",
+      "cer": 0.24485253199777407,
+      "wer": 0.5897435897435898
+    },
+    {
+      "subset_type": "accent",
+      "subset": "canada",
+      "cer": 0.18010829296095754,
+      "wer": 0.4847457627118644
+    },
+    {
+      "subset_type": "accent",
+      "subset": "burundi",
+      "cer": 0.17857142857142858,
+      "wer": 0.5714285714285714
+    },
+    {
+      "subset_type": "accent",
+      "subset": "morocco",
+      "cer": 0.0938375350140056,
+      "wer": 0.3275862068965517
+    },
+    {
+      "subset_type": "accent",
+      "subset": "belgium",
+      "cer": 0.10592389172224402,
+      "wer": 0.36596736596736595
+    },
+    {
+      "subset_type": "accent",
+      "subset": "monaco",
+      "cer": 0.1094224924012158,
+      "wer": 0.3018867924528302
+    },
+    {
+      "subset_type": "accent",
+      "subset": "senegal",
+      "cer": 0.11027190332326284,
+      "wer": 0.336283185840708
+    },
+    {
+      "subset_type": "accent",
+      "subset": "france",
+      "cer": 0.11119180678992589,
+      "wer": 0.3551616730940374
+    },
+    {
+      "subset_type": "accent",
+      "subset": "germany",
+      "cer": 0.12601078167115903,
+      "wer": 0.4048582995951417
+    },
+    {
+      "subset_type": "accent",
+      "subset": "tunisia",
+      "cer": 0.13394495412844037,
+      "wer": 0.3829787234042553
+    },
+    {
+      "subset_type": "accent",
+      "subset": "algeria",
+      "cer": 0.11131059245960502,
+      "wer": 0.40860215053763443
+    },
+    {
+      "subset_type": "accent",
+      "subset": "cameroon",
+      "cer": 0.15555555555555556,
+      "wer": 0.38961038961038963
+    },
+    {
+      "subset_type": "accent",
+      "subset": "new_caledonia",
+      "cer": 0.2140221402214022,
+      "wer": 0.4634146341463415
+    }
+  ]
+}

cv7_transcriptions.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:eb4fbdadf82a56b5abee6fcf19c9ce5ca98a7900a537b734d97d40662470775e
+size 129352303

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 sha256:092699ea8d02ae511537caac2bdbdb6a869bcc465157963a59734ecfafde48f8
+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/tmpy2fx1d9v", "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}