jonatasgrosman/exp_w2v2r_fr_xls-r_age_teens-10_sixties-0_s885

.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-10_sixties-0_s885
+
+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": 0.0756888937474992,
+      "wer": 0.2631007226883239
+    },
+    {
+      "subset_type": "age",
+      "subset": "",
+      "cer": 0.07384204885830523,
+      "wer": 0.25820164422881925
+    },
+    {
+      "subset_type": "age",
+      "subset": "twenties",
+      "cer": 0.08519543037792786,
+      "wer": 0.2924901185770751
+    },
+    {
+      "subset_type": "age",
+      "subset": "fourties",
+      "cer": 0.08209566132616068,
+      "wer": 0.2823010902342844
+    },
+    {
+      "subset_type": "age",
+      "subset": "thirties",
+      "cer": 0.08537058595265813,
+      "wer": 0.2863524337646334
+    },
+    {
+      "subset_type": "age",
+      "subset": "fifties",
+      "cer": 0.0885697684070723,
+      "wer": 0.281203007518797
+    },
+    {
+      "subset_type": "age",
+      "subset": "sixties",
+      "cer": 0.07477811167796491,
+      "wer": 0.257277397260274
+    },
+    {
+      "subset_type": "age",
+      "subset": "seventies",
+      "cer": 0.08371944400063908,
+      "wer": 0.3007662835249042
+    },
+    {
+      "subset_type": "age",
+      "subset": "teens",
+      "cer": 0.0854211255193252,
+      "wer": 0.2822018348623853
+    },
+    {
+      "subset_type": "age",
+      "subset": "eighties",
+      "cer": 0.15428571428571428,
+      "wer": 0.4482758620689655
+    },
+    {
+      "subset_type": "gender",
+      "subset": "",
+      "cer": 0.07385783587779354,
+      "wer": 0.2581945484031936
+    },
+    {
+      "subset_type": "gender",
+      "subset": "other",
+      "cer": 0.08234737340274491,
+      "wer": 0.25722543352601157
+    },
+    {
+      "subset_type": "gender",
+      "subset": "male",
+      "cer": 0.08279424909479298,
+      "wer": 0.2820755141532784
+    },
+    {
+      "subset_type": "gender",
+      "subset": "female",
+      "cer": 0.08926797920196296,
+      "wer": 0.3000352360817477
+    },
+    {
+      "subset_type": "accent",
+      "subset": "",
+      "cer": 0.07472366777750979,
+      "wer": 0.2605484187842191
+    },
+    {
+      "subset_type": "accent",
+      "subset": "cameroon",
+      "cer": 0.18585858585858586,
+      "wer": 0.5060240963855421
+    },
+    {
+      "subset_type": "accent",
+      "subset": "canada",
+      "cer": 0.12589262099973553,
+      "wer": 0.37576687116564417
+    },
+    {
+      "subset_type": "accent",
+      "subset": "guadeloupe",
+      "cer": 0.0800942285041225,
+      "wer": 0.3125
+    },
+    {
+      "subset_type": "accent",
+      "subset": "italy",
+      "cer": 0.20421052631578948,
+      "wer": 0.45569620253164556
+    },
+    {
+      "subset_type": "accent",
+      "subset": "belgium",
+      "cer": 0.08923695933323754,
+      "wer": 0.26974789915966385
+    },
+    {
+      "subset_type": "accent",
+      "subset": "luxembourg",
+      "cer": 0.08947368421052632,
+      "wer": 0.31666666666666665
+    },
+    {
+      "subset_type": "accent",
+      "subset": "romania",
+      "cer": 0.16279069767441862,
+      "wer": 0.4222222222222222
+    },
+    {
+      "subset_type": "accent",
+      "subset": "monaco",
+      "cer": 0.0273224043715847,
+      "wer": 0.15873015873015872
+    },
+    {
+      "subset_type": "accent",
+      "subset": "germany",
+      "cer": 0.05201342281879195,
+      "wer": 0.25471698113207547
+    },
+    {
+      "subset_type": "accent",
+      "subset": "united_states",
+      "cer": 0.10766045548654245,
+      "wer": 0.31076923076923074
+    },
+    {
+      "subset_type": "accent",
+      "subset": "morocco",
+      "cer": 0.06020066889632107,
+      "wer": 0.21153846153846154
+    },
+    {
+      "subset_type": "accent",
+      "subset": "martinique",
+      "cer": 0.10080645161290322,
+      "wer": 0.3157894736842105
+    },
+    {
+      "subset_type": "accent",
+      "subset": "switzerland",
+      "cer": 0.07119590507212657,
+      "wer": 0.2409972299168975
+    },
+    {
+      "subset_type": "accent",
+      "subset": "senegal",
+      "cer": 0.0979020979020979,
+      "wer": 0.36
+    },
+    {
+      "subset_type": "accent",
+      "subset": "gabon",
+      "cer": 0.1,
+      "wer": 0.3333333333333333
+    },
+    {
+      "subset_type": "accent",
+      "subset": "united_kingdom",
+      "cer": 0.0996309963099631,
+      "wer": 0.32608695652173914
+    },
+    {
+      "subset_type": "accent",
+      "subset": "mayotte",
+      "cer": 0.017167381974248927,
+      "wer": 0.10526315789473684
+    },
+    {
+      "subset_type": "accent",
+      "subset": "france",
+      "cer": 0.07508285918352607,
+      "wer": 0.2653071040353999
+    },
+    {
+      "subset_type": "accent",
+      "subset": "st_pierre_et_miquelon",
+      "cer": 0.18815331010452963,
+      "wer": 0.5833333333333334
+    },
+    {
+      "subset_type": "accent",
+      "subset": "cote_d_ivoire",
+      "cer": 0.21419518377693283,
+      "wer": 0.5909090909090909
+    },
+    {
+      "subset_type": "accent",
+      "subset": "tunisia",
+      "cer": 0.17229729729729729,
+      "wer": 0.4375
+    },
+    {
+      "subset_type": "accent",
+      "subset": "algeria",
+      "cer": 0.10586552217453506,
+      "wer": 0.344
+    }
+  ],
+  "test": [
+    {
+      "subset_type": "all",
+      "subset": "all",
+      "cer": 0.08567831207850915,
+      "wer": 0.28154213751972673
+    },
+    {
+      "subset_type": "age",
+      "subset": "",
+      "cer": 0.08466818387383838,
+      "wer": 0.27857650307242726
+    },
+    {
+      "subset_type": "age",
+      "subset": "twenties",
+      "cer": 0.09914721929301729,
+      "wer": 0.31443440153382635
+    },
+    {
+      "subset_type": "age",
+      "subset": "fourties",
+      "cer": 0.08771238888263756,
+      "wer": 0.28713527851458887
+    },
+    {
+      "subset_type": "age",
+      "subset": "thirties",
+      "cer": 0.0940703608363817,
+      "wer": 0.3075245365321701
+    },
+    {
+      "subset_type": "age",
+      "subset": "fifties",
+      "cer": 0.0848462664714495,
+      "wer": 0.2900662251655629
+    },
+    {
+      "subset_type": "age",
+      "subset": "sixties",
+      "cer": 0.08407749356742848,
+      "wer": 0.28662420382165604
+    },
+    {
+      "subset_type": "age",
+      "subset": "seventies",
+      "cer": 0.08100479819362122,
+      "wer": 0.2677966101694915
+    },
+    {
+      "subset_type": "age",
+      "subset": "teens",
+      "cer": 0.09094305783888806,
+      "wer": 0.29523809523809524
+    },
+    {
+      "subset_type": "gender",
+      "subset": "",
+      "cer": 0.08460981640991261,
+      "wer": 0.2784976841858798
+    },
+    {
+      "subset_type": "gender",
+      "subset": "other",
+      "cer": 0.14516920058852378,
+      "wer": 0.4235294117647059
+    },
+    {
+      "subset_type": "gender",
+      "subset": "male",
+      "cer": 0.09115611642580908,
+      "wer": 0.2979967119283931
+    },
+    {
+      "subset_type": "gender",
+      "subset": "female",
+      "cer": 0.09406789347353983,
+      "wer": 0.3039779259599908
+    },
+    {
+      "subset_type": "accent",
+      "subset": "",
+      "cer": 0.08550732524464803,
+      "wer": 0.28071612783722555
+    },
+    {
+      "subset_type": "accent",
+      "subset": "other",
+      "cer": 0.037383177570093455,
+      "wer": 0.25
+    },
+    {
+      "subset_type": "accent",
+      "subset": "cameroon",
+      "cer": 0.13555555555555557,
+      "wer": 0.36363636363636365
+    },
+    {
+      "subset_type": "accent",
+      "subset": "canada",
+      "cer": 0.14120832145910517,
+      "wer": 0.40084745762711865
+    },
+    {
+      "subset_type": "accent",
+      "subset": "reunion",
+      "cer": 0.06269592476489028,
+      "wer": 0.2037037037037037
+    },
+    {
+      "subset_type": "accent",
+      "subset": "guadeloupe",
+      "cer": 0.015810276679841896,
+      "wer": 0.09523809523809523
+    },
+    {
+      "subset_type": "accent",
+      "subset": "benin",
+      "cer": 0.12727272727272726,
+      "wer": 0.34210526315789475
+    },
+    {
+      "subset_type": "accent",
+      "subset": "belgium",
+      "cer": 0.07630443311102393,
+      "wer": 0.28088578088578087
+    },
+    {
+      "subset_type": "accent",
+      "subset": "romania",
+      "cer": 0.0916030534351145,
+      "wer": 0.3181818181818182
+    },
+    {
+      "subset_type": "accent",
+      "subset": "monaco",
+      "cer": 0.10638297872340426,
+      "wer": 0.2830188679245283
+    },
+    {
+      "subset_type": "accent",
+      "subset": "germany",
+      "cer": 0.10646900269541779,
+      "wer": 0.3319838056680162
+    },
+    {
+      "subset_type": "accent",
+      "subset": "united_states",
+      "cer": 0.21702838063439064,
+      "wer": 0.5448717948717948
+    },
+    {
+      "subset_type": "accent",
+      "subset": "morocco",
+      "cer": 0.0784313725490196,
+      "wer": 0.25862068965517243
+    },
+    {
+      "subset_type": "accent",
+      "subset": "martinique",
+      "cer": 0.04824561403508772,
+      "wer": 0.21052631578947367
+    },
+    {
+      "subset_type": "accent",
+      "subset": "burundi",
+      "cer": 0.11904761904761904,
+      "wer": 0.35714285714285715
+    },
+    {
+      "subset_type": "accent",
+      "subset": "switzerland",
+      "cer": 0.0815450643776824,
+      "wer": 0.2737226277372263
+    },
+    {
+      "subset_type": "accent",
+      "subset": "senegal",
+      "cer": 0.11329305135951662,
+      "wer": 0.336283185840708
+    },
+    {
+      "subset_type": "accent",
+      "subset": "united_kingdom",
+      "cer": 0.17142857142857143,
+      "wer": 0.3333333333333333
+    },
+    {
+      "subset_type": "accent",
+      "subset": "france",
+      "cer": 0.07854698540092075,
+      "wer": 0.2700237318303174
+    },
+    {
+      "subset_type": "accent",
+      "subset": "cote_d_ivoire",
+      "cer": 0.14009111617312073,
+      "wer": 0.4125874125874126
+    },
+    {
+      "subset_type": "accent",
+      "subset": "new_caledonia",
+      "cer": 0.1070110701107011,
+      "wer": 0.34146341463414637
+    },
+    {
+      "subset_type": "accent",
+      "subset": "tunisia",
+      "cer": 0.11009174311926606,
+      "wer": 0.2978723404255319
+    },
+    {
+      "subset_type": "accent",
+      "subset": "algeria",
+      "cer": 0.11759425493716337,
+      "wer": 0.3709677419354839
+    }
+  ]
+}

cv7_transcriptions.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0b79aa98411029f1aca7f9b52f42f972b8fd305b892e19da55f3a95c61a213fb
+size 130576672

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 sha256:63c806322701d6dfe62df981c2c73a8eaff1f3a7ff23026b521610d5b6443bce
+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/tmpiv_b2cmf", "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}