jonatasgrosman/exp_w2v2r_fr_xls-r_age_teens-0_sixties-10_s888

.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-0_sixties-10_s888
+
+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": 0.0822310361186968,
+      "wer": 0.2847813724049355
+    },
+    {
+      "subset_type": "age",
+      "subset": "",
+      "cer": 0.07997357612957522,
+      "wer": 0.2791133020295462
+    },
+    {
+      "subset_type": "age",
+      "subset": "twenties",
+      "cer": 0.09868515591320592,
+      "wer": 0.32443115051810706
+    },
+    {
+      "subset_type": "age",
+      "subset": "fourties",
+      "cer": 0.08985303863095934,
+      "wer": 0.3119925771282765
+    },
+    {
+      "subset_type": "age",
+      "subset": "thirties",
+      "cer": 0.09341611693183288,
+      "wer": 0.3139248305606901
+    },
+    {
+      "subset_type": "age",
+      "subset": "fifties",
+      "cer": 0.08964887523864863,
+      "wer": 0.293734335839599
+    },
+    {
+      "subset_type": "age",
+      "subset": "sixties",
+      "cer": 0.0701656300230624,
+      "wer": 0.2538527397260274
+    },
+    {
+      "subset_type": "age",
+      "subset": "seventies",
+      "cer": 0.07892634606167119,
+      "wer": 0.27490421455938696
+    },
+    {
+      "subset_type": "age",
+      "subset": "teens",
+      "cer": 0.10115825254941457,
+      "wer": 0.32513761467889907
+    },
+    {
+      "subset_type": "age",
+      "subset": "eighties",
+      "cer": 0.11428571428571428,
+      "wer": 0.41379310344827586
+    },
+    {
+      "subset_type": "gender",
+      "subset": "",
+      "cer": 0.07996634954602902,
+      "wer": 0.27912924151696605
+    },
+    {
+      "subset_type": "gender",
+      "subset": "other",
+      "cer": 0.09985802176999527,
+      "wer": 0.3208092485549133
+    },
+    {
+      "subset_type": "gender",
+      "subset": "male",
+      "cer": 0.09042980289969246,
+      "wer": 0.3043967418208001
+    },
+    {
+      "subset_type": "gender",
+      "subset": "female",
+      "cer": 0.10071858386399486,
+      "wer": 0.3335095137420719
+    },
+    {
+      "subset_type": "accent",
+      "subset": "",
+      "cer": 0.08079142205177228,
+      "wer": 0.2814967675121647
+    },
+    {
+      "subset_type": "accent",
+      "subset": "cameroon",
+      "cer": 0.1898989898989899,
+      "wer": 0.5060240963855421
+    },
+    {
+      "subset_type": "accent",
+      "subset": "canada",
+      "cer": 0.14850568632636868,
+      "wer": 0.42714723926380366
+    },
+    {
+      "subset_type": "accent",
+      "subset": "guadeloupe",
+      "cer": 0.0977620730270907,
+      "wer": 0.3263888888888889
+    },
+    {
+      "subset_type": "accent",
+      "subset": "italy",
+      "cer": 0.21263157894736842,
+      "wer": 0.620253164556962
+    },
+    {
+      "subset_type": "accent",
+      "subset": "belgium",
+      "cer": 0.09254203190113522,
+      "wer": 0.2848739495798319
+    },
+    {
+      "subset_type": "accent",
+      "subset": "luxembourg",
+      "cer": 0.14736842105263157,
+      "wer": 0.43333333333333335
+    },
+    {
+      "subset_type": "accent",
+      "subset": "romania",
+      "cer": 0.13953488372093023,
+      "wer": 0.4666666666666667
+    },
+    {
+      "subset_type": "accent",
+      "subset": "monaco",
+      "cer": 0.030054644808743168,
+      "wer": 0.1746031746031746
+    },
+    {
+      "subset_type": "accent",
+      "subset": "germany",
+      "cer": 0.0738255033557047,
+      "wer": 0.32075471698113206
+    },
+    {
+      "subset_type": "accent",
+      "subset": "united_states",
+      "cer": 0.12732919254658384,
+      "wer": 0.36923076923076925
+    },
+    {
+      "subset_type": "accent",
+      "subset": "morocco",
+      "cer": 0.056856187290969896,
+      "wer": 0.19230769230769232
+    },
+    {
+      "subset_type": "accent",
+      "subset": "martinique",
+      "cer": 0.0846774193548387,
+      "wer": 0.3157894736842105
+    },
+    {
+      "subset_type": "accent",
+      "subset": "switzerland",
+      "cer": 0.08469055374592833,
+      "wer": 0.27146814404432135
+    },
+    {
+      "subset_type": "accent",
+      "subset": "senegal",
+      "cer": 0.11538461538461539,
+      "wer": 0.46
+    },
+    {
+      "subset_type": "accent",
+      "subset": "gabon",
+      "cer": 0.08888888888888889,
+      "wer": 0.3541666666666667
+    },
+    {
+      "subset_type": "accent",
+      "subset": "united_kingdom",
+      "cer": 0.11070110701107011,
+      "wer": 0.30434782608695654
+    },
+    {
+      "subset_type": "accent",
+      "subset": "mayotte",
+      "cer": 0.017167381974248927,
+      "wer": 0.07894736842105263
+    },
+    {
+      "subset_type": "accent",
+      "subset": "france",
+      "cer": 0.08306783619160735,
+      "wer": 0.28863450531479967
+    },
+    {
+      "subset_type": "accent",
+      "subset": "st_pierre_et_miquelon",
+      "cer": 0.18815331010452963,
+      "wer": 0.625
+    },
+    {
+      "subset_type": "accent",
+      "subset": "cote_d_ivoire",
+      "cer": 0.2674271229404309,
+      "wer": 0.6439393939393939
+    },
+    {
+      "subset_type": "accent",
+      "subset": "tunisia",
+      "cer": 0.2195945945945946,
+      "wer": 0.5
+    },
+    {
+      "subset_type": "accent",
+      "subset": "algeria",
+      "cer": 0.10014306151645208,
+      "wer": 0.304
+    }
+  ],
+  "test": [
+    {
+      "subset_type": "all",
+      "subset": "all",
+      "cer": 0.09364018604024416,
+      "wer": 0.30572755908506094
+    },
+    {
+      "subset_type": "age",
+      "subset": "",
+      "cer": 0.09229555390218877,
+      "wer": 0.3018302731127826
+    },
+    {
+      "subset_type": "age",
+      "subset": "twenties",
+      "cer": 0.11542341020585943,
+      "wer": 0.3601752944398795
+    },
+    {
+      "subset_type": "age",
+      "subset": "fourties",
+      "cer": 0.09255091706987735,
+      "wer": 0.3116710875331565
+    },
+    {
+      "subset_type": "age",
+      "subset": "thirties",
+      "cer": 0.10423158560267354,
+      "wer": 0.3356052344601963
+    },
+    {
+      "subset_type": "age",
+      "subset": "fifties",
+      "cer": 0.08513909224011713,
+      "wer": 0.2883002207505519
+    },
+    {
+      "subset_type": "age",
+      "subset": "sixties",
+      "cer": 0.08710458604510368,
+      "wer": 0.29572338489535943
+    },
+    {
+      "subset_type": "age",
+      "subset": "seventies",
+      "cer": 0.08156929156082417,
+      "wer": 0.27288135593220336
+    },
+    {
+      "subset_type": "age",
+      "subset": "teens",
+      "cer": 0.10753250635181587,
+      "wer": 0.345578231292517
+    },
+    {
+      "subset_type": "gender",
+      "subset": "",
+      "cer": 0.09222013321090755,
+      "wer": 0.3017143609743248
+    },
+    {
+      "subset_type": "gender",
+      "subset": "other",
+      "cer": 0.15890142226581658,
+      "wer": 0.4264705882352941
+    },
+    {
+      "subset_type": "gender",
+      "subset": "male",
+      "cer": 0.10035619784245878,
+      "wer": 0.3266760031662912
+    },
+    {
+      "subset_type": "gender",
+      "subset": "female",
+      "cer": 0.10797424467558198,
+      "wer": 0.3432973097263739
+    },
+    {
+      "subset_type": "accent",
+      "subset": "",
+      "cer": 0.09322296568302701,
+      "wer": 0.3042545915677586
+    },
+    {
+      "subset_type": "accent",
+      "subset": "other",
+      "cer": 0.037383177570093455,
+      "wer": 0.15
+    },
+    {
+      "subset_type": "accent",
+      "subset": "cameroon",
+      "cer": 0.14444444444444443,
+      "wer": 0.4025974025974026
+    },
+    {
+      "subset_type": "accent",
+      "subset": "canada",
+      "cer": 0.1487603305785124,
+      "wer": 0.4211864406779661
+    },
+    {
+      "subset_type": "accent",
+      "subset": "reunion",
+      "cer": 0.08463949843260188,
+      "wer": 0.24074074074074073
+    },
+    {
+      "subset_type": "accent",
+      "subset": "guadeloupe",
+      "cer": 0.015810276679841896,
+      "wer": 0.07142857142857142
+    },
+    {
+      "subset_type": "accent",
+      "subset": "benin",
+      "cer": 0.07727272727272727,
+      "wer": 0.3157894736842105
+    },
+    {
+      "subset_type": "accent",
+      "subset": "belgium",
+      "cer": 0.07787367595135347,
+      "wer": 0.289044289044289
+    },
+    {
+      "subset_type": "accent",
+      "subset": "romania",
+      "cer": 0.12213740458015267,
+      "wer": 0.3181818181818182
+    },
+    {
+      "subset_type": "accent",
+      "subset": "monaco",
+      "cer": 0.11854103343465046,
+      "wer": 0.33962264150943394
+    },
+    {
+      "subset_type": "accent",
+      "subset": "germany",
+      "cer": 0.11725067385444744,
+      "wer": 0.3805668016194332
+    },
+    {
+      "subset_type": "accent",
+      "subset": "united_states",
+      "cer": 0.28213689482470783,
+      "wer": 0.6698717948717948
+    },
+    {
+      "subset_type": "accent",
+      "subset": "morocco",
+      "cer": 0.07563025210084033,
+      "wer": 0.25862068965517243
+    },
+    {
+      "subset_type": "accent",
+      "subset": "martinique",
+      "cer": 0.10087719298245613,
+      "wer": 0.2631578947368421
+    },
+    {
+      "subset_type": "accent",
+      "subset": "burundi",
+      "cer": 0.13095238095238096,
+      "wer": 0.5714285714285714
+    },
+    {
+      "subset_type": "accent",
+      "subset": "switzerland",
+      "cer": 0.07909258123850399,
+      "wer": 0.30656934306569344
+    },
+    {
+      "subset_type": "accent",
+      "subset": "senegal",
+      "cer": 0.11178247734138973,
+      "wer": 0.34513274336283184
+    },
+    {
+      "subset_type": "accent",
+      "subset": "united_kingdom",
+      "cer": 0.3142857142857143,
+      "wer": 0.6666666666666666
+    },
+    {
+      "subset_type": "accent",
+      "subset": "france",
+      "cer": 0.0886777124006007,
+      "wer": 0.299614357757342
+    },
+    {
+      "subset_type": "accent",
+      "subset": "cote_d_ivoire",
+      "cer": 0.13325740318906606,
+      "wer": 0.45454545454545453
+    },
+    {
+      "subset_type": "accent",
+      "subset": "new_caledonia",
+      "cer": 0.09225092250922509,
+      "wer": 0.34146341463414637
+    },
+    {
+      "subset_type": "accent",
+      "subset": "tunisia",
+      "cer": 0.14862385321100918,
+      "wer": 0.3829787234042553
+    },
+    {
+      "subset_type": "accent",
+      "subset": "algeria",
+      "cer": 0.11220825852782765,
+      "wer": 0.34408602150537637
+    }
+  ]
+}

cv7_transcriptions.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4eecfb9a3c252a47ae99046a8cedc7f6df96643bf072ef23d6d294f6dfccd968
+size 130152628

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:ce71f1d0443d9e7bc975211cde4621b4448296b4f47e2ebdf686d4ebb5390384
+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/tmpndxdbkw3", "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}