jonatasgrosman/exp_w2v2r_fr_xls-r_age_teens-5_sixties-5_s362

.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-5_sixties-5_s362
+
+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": 0.07656025956300103,
+      "wer": 0.26795059393350756
+    },
+    {
+      "subset_type": "age",
+      "subset": "",
+      "cer": 0.07454789197239196,
+      "wer": 0.2628114128823924
+    },
+    {
+      "subset_type": "age",
+      "subset": "twenties",
+      "cer": 0.08916510992958758,
+      "wer": 0.29911334259160344
+    },
+    {
+      "subset_type": "age",
+      "subset": "fourties",
+      "cer": 0.08498031419327175,
+      "wer": 0.2964509394572025
+    },
+    {
+      "subset_type": "age",
+      "subset": "thirties",
+      "cer": 0.08568102444703143,
+      "wer": 0.29066543438077636
+    },
+    {
+      "subset_type": "age",
+      "subset": "fifties",
+      "cer": 0.08549846434797045,
+      "wer": 0.27268170426065164
+    },
+    {
+      "subset_type": "age",
+      "subset": "sixties",
+      "cer": 0.07296107345027605,
+      "wer": 0.257277397260274
+    },
+    {
+      "subset_type": "age",
+      "subset": "seventies",
+      "cer": 0.07796772647387762,
+      "wer": 0.2835249042145594
+    },
+    {
+      "subset_type": "age",
+      "subset": "teens",
+      "cer": 0.09026816064459273,
+      "wer": 0.30275229357798167
+    },
+    {
+      "subset_type": "age",
+      "subset": "eighties",
+      "cer": 0.09714285714285714,
+      "wer": 0.3103448275862069
+    },
+    {
+      "subset_type": "gender",
+      "subset": "",
+      "cer": 0.07455565635184626,
+      "wer": 0.2628103168662675
+    },
+    {
+      "subset_type": "gender",
+      "subset": "other",
+      "cer": 0.07903454803596782,
+      "wer": 0.2745664739884393
+    },
+    {
+      "subset_type": "gender",
+      "subset": "male",
+      "cer": 0.08393049222052207,
+      "wer": 0.2864407542414833
+    },
+    {
+      "subset_type": "gender",
+      "subset": "female",
+      "cer": 0.09329905941461705,
+      "wer": 0.31131078224101477
+    },
+    {
+      "subset_type": "accent",
+      "subset": "",
+      "cer": 0.0753765354070194,
+      "wer": 0.26497262474966304
+    },
+    {
+      "subset_type": "accent",
+      "subset": "cameroon",
+      "cer": 0.2,
+      "wer": 0.4819277108433735
+    },
+    {
+      "subset_type": "accent",
+      "subset": "canada",
+      "cer": 0.12377677862999206,
+      "wer": 0.36119631901840493
+    },
+    {
+      "subset_type": "accent",
+      "subset": "guadeloupe",
+      "cer": 0.08480565371024736,
+      "wer": 0.3263888888888889
+    },
+    {
+      "subset_type": "accent",
+      "subset": "italy",
+      "cer": 0.16210526315789472,
+      "wer": 0.4810126582278481
+    },
+    {
+      "subset_type": "accent",
+      "subset": "belgium",
+      "cer": 0.08765627245293864,
+      "wer": 0.27647058823529413
+    },
+    {
+      "subset_type": "accent",
+      "subset": "luxembourg",
+      "cer": 0.14210526315789473,
+      "wer": 0.4166666666666667
+    },
+    {
+      "subset_type": "accent",
+      "subset": "romania",
+      "cer": 0.14728682170542637,
+      "wer": 0.5111111111111111
+    },
+    {
+      "subset_type": "accent",
+      "subset": "monaco",
+      "cer": 0.02459016393442623,
+      "wer": 0.15873015873015872
+    },
+    {
+      "subset_type": "accent",
+      "subset": "germany",
+      "cer": 0.07046979865771812,
+      "wer": 0.3018867924528302
+    },
+    {
+      "subset_type": "accent",
+      "subset": "united_states",
+      "cer": 0.10507246376811594,
+      "wer": 0.3230769230769231
+    },
+    {
+      "subset_type": "accent",
+      "subset": "morocco",
+      "cer": 0.033444816053511704,
+      "wer": 0.11538461538461539
+    },
+    {
+      "subset_type": "accent",
+      "subset": "martinique",
+      "cer": 0.09274193548387097,
+      "wer": 0.39473684210526316
+    },
+    {
+      "subset_type": "accent",
+      "subset": "switzerland",
+      "cer": 0.07864122847836202,
+      "wer": 0.25761772853185594
+    },
+    {
+      "subset_type": "accent",
+      "subset": "senegal",
+      "cer": 0.11188811188811189,
+      "wer": 0.36
+    },
+    {
+      "subset_type": "accent",
+      "subset": "gabon",
+      "cer": 0.06296296296296296,
+      "wer": 0.2708333333333333
+    },
+    {
+      "subset_type": "accent",
+      "subset": "united_kingdom",
+      "cer": 0.13284132841328414,
+      "wer": 0.34782608695652173
+    },
+    {
+      "subset_type": "accent",
+      "subset": "mayotte",
+      "cer": 0.004291845493562232,
+      "wer": 0.02631578947368421
+    },
+    {
+      "subset_type": "accent",
+      "subset": "france",
+      "cer": 0.0773780385044419,
+      "wer": 0.2731470347746621
+    },
+    {
+      "subset_type": "accent",
+      "subset": "st_pierre_et_miquelon",
+      "cer": 0.18466898954703834,
+      "wer": 0.625
+    },
+    {
+      "subset_type": "accent",
+      "subset": "cote_d_ivoire",
+      "cer": 0.24461343472750316,
+      "wer": 0.6363636363636364
+    },
+    {
+      "subset_type": "accent",
+      "subset": "tunisia",
+      "cer": 0.19932432432432431,
+      "wer": 0.4375
+    },
+    {
+      "subset_type": "accent",
+      "subset": "algeria",
+      "cer": 0.10014306151645208,
+      "wer": 0.328
+    }
+  ],
+  "test": [
+    {
+      "subset_type": "all",
+      "subset": "all",
+      "cer": 0.08807540146246666,
+      "wer": 0.28872297678427694
+    },
+    {
+      "subset_type": "age",
+      "subset": "",
+      "cer": 0.08693062181179823,
+      "wer": 0.28581263686766223
+    },
+    {
+      "subset_type": "age",
+      "subset": "twenties",
+      "cer": 0.10288386593920498,
+      "wer": 0.32593809915091754
+    },
+    {
+      "subset_type": "age",
+      "subset": "fourties",
+      "cer": 0.09198829751322156,
+      "wer": 0.28713527851458887
+    },
+    {
+      "subset_type": "age",
+      "subset": "thirties",
+      "cer": 0.09967032470758252,
+      "wer": 0.32006543075245364
+    },
+    {
+      "subset_type": "age",
+      "subset": "fifties",
+      "cer": 0.07994143484626647,
+      "wer": 0.2772626931567329
+    },
+    {
+      "subset_type": "age",
+      "subset": "sixties",
+      "cer": 0.08604510367791736,
+      "wer": 0.28343949044585987
+    },
+    {
+      "subset_type": "age",
+      "subset": "seventies",
+      "cer": 0.07705334462320068,
+      "wer": 0.27288135593220336
+    },
+    {
+      "subset_type": "age",
+      "subset": "teens",
+      "cer": 0.09901360035869078,
+      "wer": 0.31700680272108844
+    },
+    {
+      "subset_type": "gender",
+      "subset": "",
+      "cer": 0.08686279082195308,
+      "wer": 0.2857121951576436
+    },
+    {
+      "subset_type": "gender",
+      "subset": "other",
+      "cer": 0.15841098577734183,
+      "wer": 0.40588235294117647
+    },
+    {
+      "subset_type": "gender",
+      "subset": "male",
+      "cer": 0.09422959495216772,
+      "wer": 0.3039639529927541
+    },
+    {
+      "subset_type": "gender",
+      "subset": "female",
+      "cer": 0.0976111555606355,
+      "wer": 0.3166245113819269
+    },
+    {
+      "subset_type": "accent",
+      "subset": "",
+      "cer": 0.08775150953038935,
+      "wer": 0.2878261117364604
+    },
+    {
+      "subset_type": "accent",
+      "subset": "other",
+      "cer": 0.018691588785046728,
+      "wer": 0.15
+    },
+    {
+      "subset_type": "accent",
+      "subset": "cameroon",
+      "cer": 0.14444444444444443,
+      "wer": 0.42857142857142855
+    },
+    {
+      "subset_type": "accent",
+      "subset": "canada",
+      "cer": 0.14291821031632942,
+      "wer": 0.41271186440677965
+    },
+    {
+      "subset_type": "accent",
+      "subset": "reunion",
+      "cer": 0.08463949843260188,
+      "wer": 0.2222222222222222
+    },
+    {
+      "subset_type": "accent",
+      "subset": "guadeloupe",
+      "cer": 0.015810276679841896,
+      "wer": 0.09523809523809523
+    },
+    {
+      "subset_type": "accent",
+      "subset": "benin",
+      "cer": 0.09090909090909091,
+      "wer": 0.34210526315789475
+    },
+    {
+      "subset_type": "accent",
+      "subset": "belgium",
+      "cer": 0.07414672420557081,
+      "wer": 0.25524475524475526
+    },
+    {
+      "subset_type": "accent",
+      "subset": "romania",
+      "cer": 0.1450381679389313,
+      "wer": 0.36363636363636365
+    },
+    {
+      "subset_type": "accent",
+      "subset": "monaco",
+      "cer": 0.0851063829787234,
+      "wer": 0.24528301886792453
+    },
+    {
+      "subset_type": "accent",
+      "subset": "germany",
+      "cer": 0.10040431266846361,
+      "wer": 0.3481781376518219
+    },
+    {
+      "subset_type": "accent",
+      "subset": "united_states",
+      "cer": 0.23094045631608237,
+      "wer": 0.5705128205128205
+    },
+    {
+      "subset_type": "accent",
+      "subset": "morocco",
+      "cer": 0.08123249299719888,
+      "wer": 0.27586206896551724
+    },
+    {
+      "subset_type": "accent",
+      "subset": "martinique",
+      "cer": 0.02631578947368421,
+      "wer": 0.15789473684210525
+    },
+    {
+      "subset_type": "accent",
+      "subset": "burundi",
+      "cer": 0.16666666666666666,
+      "wer": 0.5714285714285714
+    },
+    {
+      "subset_type": "accent",
+      "subset": "switzerland",
+      "cer": 0.06928264868179031,
+      "wer": 0.2773722627737226
+    },
+    {
+      "subset_type": "accent",
+      "subset": "senegal",
+      "cer": 0.09365558912386707,
+      "wer": 0.2831858407079646
+    },
+    {
+      "subset_type": "accent",
+      "subset": "united_kingdom",
+      "cer": 0.2857142857142857,
+      "wer": 0.5555555555555556
+    },
+    {
+      "subset_type": "accent",
+      "subset": "france",
+      "cer": 0.0832492183460942,
+      "wer": 0.2791456541085731
+    },
+    {
+      "subset_type": "accent",
+      "subset": "cote_d_ivoire",
+      "cer": 0.14009111617312073,
+      "wer": 0.42657342657342656
+    },
+    {
+      "subset_type": "accent",
+      "subset": "new_caledonia",
+      "cer": 0.0996309963099631,
+      "wer": 0.2682926829268293
+    },
+    {
+      "subset_type": "accent",
+      "subset": "tunisia",
+      "cer": 0.13211009174311927,
+      "wer": 0.39361702127659576
+    },
+    {
+      "subset_type": "accent",
+      "subset": "algeria",
+      "cer": 0.11579892280071813,
+      "wer": 0.3548387096774194
+    }
+  ]
+}

cv7_transcriptions.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3f14d0acfb6301164f8d8919738b64e2603d9ba7e2915d6e0162cba65cd28831
+size 130453691

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:cb6a6ec0059cdfdcdbaa96adb18df74ea647d89dabde5fbee2c7324fc514ea94
+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/tmp99itsop6", "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}