jonatasgrosman/exp_w2v2r_fr_vp-100k_age_teens-0_sixties-10_s423

.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-0_sixties-10_s423
+
+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": 0.10543844771645905,
+      "wer": 0.346372820273612
+    },
+    {
+      "subset_type": "age",
+      "subset": "",
+      "cer": 0.10283461940706592,
+      "wer": 0.3408108824859991
+    },
+    {
+      "subset_type": "age",
+      "subset": "thirties",
+      "cer": 0.11840641572888372,
+      "wer": 0.37153419593345655
+    },
+    {
+      "subset_type": "age",
+      "subset": "sixties",
+      "cer": 0.09567405129638688,
+      "wer": 0.3214897260273973
+    },
+    {
+      "subset_type": "age",
+      "subset": "seventies",
+      "cer": 0.10369068541300527,
+      "wer": 0.3448275862068966
+    },
+    {
+      "subset_type": "age",
+      "subset": "fifties",
+      "cer": 0.1204449240474807,
+      "wer": 0.3649122807017544
+    },
+    {
+      "subset_type": "age",
+      "subset": "twenties",
+      "cer": 0.12198232504670212,
+      "wer": 0.3814763379980771
+    },
+    {
+      "subset_type": "age",
+      "subset": "eighties",
+      "cer": 0.08,
+      "wer": 0.3448275862068966
+    },
+    {
+      "subset_type": "age",
+      "subset": "teens",
+      "cer": 0.1226236938184565,
+      "wer": 0.3853211009174312
+    },
+    {
+      "subset_type": "age",
+      "subset": "fourties",
+      "cer": 0.1168674229134994,
+      "wer": 0.37833449315704015
+    },
+    {
+      "subset_type": "gender",
+      "subset": "",
+      "cer": 0.10283418493276371,
+      "wer": 0.34080276946107785
+    },
+    {
+      "subset_type": "gender",
+      "subset": "male",
+      "cer": 0.11501810414046995,
+      "wer": 0.3687952837406057
+    },
+    {
+      "subset_type": "gender",
+      "subset": "other",
+      "cer": 0.110269758637009,
+      "wer": 0.34104046242774566
+    },
+    {
+      "subset_type": "gender",
+      "subset": "female",
+      "cer": 0.1270666588771397,
+      "wer": 0.38477801268498946
+    },
+    {
+      "subset_type": "accent",
+      "subset": "",
+      "cer": 0.1039107654967666,
+      "wer": 0.34341280659138157
+    },
+    {
+      "subset_type": "accent",
+      "subset": "united_kingdom",
+      "cer": 0.13284132841328414,
+      "wer": 0.4782608695652174
+    },
+    {
+      "subset_type": "accent",
+      "subset": "st_pierre_et_miquelon",
+      "cer": 0.30313588850174217,
+      "wer": 0.6875
+    },
+    {
+      "subset_type": "accent",
+      "subset": "martinique",
+      "cer": 0.1814516129032258,
+      "wer": 0.47368421052631576
+    },
+    {
+      "subset_type": "accent",
+      "subset": "romania",
+      "cer": 0.18604651162790697,
+      "wer": 0.6222222222222222
+    },
+    {
+      "subset_type": "accent",
+      "subset": "cote_d_ivoire",
+      "cer": 0.30164765525982257,
+      "wer": 0.6666666666666666
+    },
+    {
+      "subset_type": "accent",
+      "subset": "guadeloupe",
+      "cer": 0.0989399293286219,
+      "wer": 0.375
+    },
+    {
+      "subset_type": "accent",
+      "subset": "switzerland",
+      "cer": 0.09632387156817124,
+      "wer": 0.3157894736842105
+    },
+    {
+      "subset_type": "accent",
+      "subset": "united_states",
+      "cer": 0.14699792960662525,
+      "wer": 0.37846153846153846
+    },
+    {
+      "subset_type": "accent",
+      "subset": "canada",
+      "cer": 0.17098651150489289,
+      "wer": 0.4938650306748466
+    },
+    {
+      "subset_type": "accent",
+      "subset": "morocco",
+      "cer": 0.05016722408026756,
+      "wer": 0.19230769230769232
+    },
+    {
+      "subset_type": "accent",
+      "subset": "belgium",
+      "cer": 0.11596493749101883,
+      "wer": 0.3521008403361345
+    },
+    {
+      "subset_type": "accent",
+      "subset": "luxembourg",
+      "cer": 0.1368421052631579,
+      "wer": 0.45
+    },
+    {
+      "subset_type": "accent",
+      "subset": "italy",
+      "cer": 0.21263157894736842,
+      "wer": 0.5443037974683544
+    },
+    {
+      "subset_type": "accent",
+      "subset": "monaco",
+      "cer": 0.060109289617486336,
+      "wer": 0.3333333333333333
+    },
+    {
+      "subset_type": "accent",
+      "subset": "senegal",
+      "cer": 0.19230769230769232,
+      "wer": 0.54
+    },
+    {
+      "subset_type": "accent",
+      "subset": "mayotte",
+      "cer": 0.03862660944206009,
+      "wer": 0.21052631578947367
+    },
+    {
+      "subset_type": "accent",
+      "subset": "france",
+      "cer": 0.10629248288646888,
+      "wer": 0.3480833052763215
+    },
+    {
+      "subset_type": "accent",
+      "subset": "germany",
+      "cer": 0.11409395973154363,
+      "wer": 0.4056603773584906
+    },
+    {
+      "subset_type": "accent",
+      "subset": "tunisia",
+      "cer": 0.27364864864864863,
+      "wer": 0.5208333333333334
+    },
+    {
+      "subset_type": "accent",
+      "subset": "gabon",
+      "cer": 0.1111111111111111,
+      "wer": 0.3125
+    },
+    {
+      "subset_type": "accent",
+      "subset": "algeria",
+      "cer": 0.17310443490701002,
+      "wer": 0.496
+    },
+    {
+      "subset_type": "accent",
+      "subset": "cameroon",
+      "cer": 0.26262626262626265,
+      "wer": 0.5783132530120482
+    }
+  ],
+  "test": [
+    {
+      "subset_type": "all",
+      "subset": "all",
+      "cer": 0.12038083928657904,
+      "wer": 0.37024103028881805
+    },
+    {
+      "subset_type": "age",
+      "subset": "",
+      "cer": 0.11920774553712538,
+      "wer": 0.366721108563979
+    },
+    {
+      "subset_type": "age",
+      "subset": "thirties",
+      "cer": 0.13281849794517456,
+      "wer": 0.40294438386041437
+    },
+    {
+      "subset_type": "age",
+      "subset": "seventies",
+      "cer": 0.09935083262771663,
+      "wer": 0.31864406779661014
+    },
+    {
+      "subset_type": "age",
+      "subset": "fifties",
+      "cer": 0.11251830161054173,
+      "wer": 0.35320088300220753
+    },
+    {
+      "subset_type": "age",
+      "subset": "twenties",
+      "cer": 0.13573426252808216,
+      "wer": 0.4150917556833744
+    },
+    {
+      "subset_type": "age",
+      "subset": "fourties",
+      "cer": 0.12512658940024754,
+      "wer": 0.3796419098143236
+    },
+    {
+      "subset_type": "age",
+      "subset": "teens",
+      "cer": 0.13279031534897623,
+      "wer": 0.4054421768707483
+    },
+    {
+      "subset_type": "age",
+      "subset": "sixties",
+      "cer": 0.11669441501437869,
+      "wer": 0.3685168334849864
+    },
+    {
+      "subset_type": "gender",
+      "subset": "",
+      "cer": 0.11910279393117923,
+      "wer": 0.36660837497896376
+    },
+    {
+      "subset_type": "gender",
+      "subset": "male",
+      "cer": 0.1256971300630979,
+      "wer": 0.38573951166047615
+    },
+    {
+      "subset_type": "gender",
+      "subset": "other",
+      "cer": 0.20941638057871506,
+      "wer": 0.5147058823529411
+    },
+    {
+      "subset_type": "gender",
+      "subset": "female",
+      "cer": 0.13365336990894197,
+      "wer": 0.4145780639227409
+    },
+    {
+      "subset_type": "accent",
+      "subset": "",
+      "cer": 0.12000175086596884,
+      "wer": 0.3690281122209082
+    },
+    {
+      "subset_type": "accent",
+      "subset": "united_kingdom",
+      "cer": 0.2857142857142857,
+      "wer": 0.5555555555555556
+    },
+    {
+      "subset_type": "accent",
+      "subset": "benin",
+      "cer": 0.09090909090909091,
+      "wer": 0.34210526315789475
+    },
+    {
+      "subset_type": "accent",
+      "subset": "reunion",
+      "cer": 0.11285266457680251,
+      "wer": 0.4074074074074074
+    },
+    {
+      "subset_type": "accent",
+      "subset": "martinique",
+      "cer": 0.09649122807017543,
+      "wer": 0.23684210526315788
+    },
+    {
+      "subset_type": "accent",
+      "subset": "romania",
+      "cer": 0.16793893129770993,
+      "wer": 0.36363636363636365
+    },
+    {
+      "subset_type": "accent",
+      "subset": "cote_d_ivoire",
+      "cer": 0.21753986332574032,
+      "wer": 0.5804195804195804
+    },
+    {
+      "subset_type": "accent",
+      "subset": "other",
+      "cer": 0.04672897196261682,
+      "wer": 0.2
+    },
+    {
+      "subset_type": "accent",
+      "subset": "guadeloupe",
+      "cer": 0.05138339920948617,
+      "wer": 0.23809523809523808
+    },
+    {
+      "subset_type": "accent",
+      "subset": "switzerland",
+      "cer": 0.11710606989576947,
+      "wer": 0.35036496350364965
+    },
+    {
+      "subset_type": "accent",
+      "subset": "united_states",
+      "cer": 0.2609905397885364,
+      "wer": 0.6282051282051282
+    },
+    {
+      "subset_type": "accent",
+      "subset": "canada",
+      "cer": 0.1786833855799373,
+      "wer": 0.48389830508474574
+    },
+    {
+      "subset_type": "accent",
+      "subset": "burundi",
+      "cer": 0.21428571428571427,
+      "wer": 0.7142857142857143
+    },
+    {
+      "subset_type": "accent",
+      "subset": "morocco",
+      "cer": 0.08683473389355742,
+      "wer": 0.28448275862068967
+    },
+    {
+      "subset_type": "accent",
+      "subset": "belgium",
+      "cer": 0.1037661828167909,
+      "wer": 0.34382284382284384
+    },
+    {
+      "subset_type": "accent",
+      "subset": "monaco",
+      "cer": 0.11854103343465046,
+      "wer": 0.3584905660377358
+    },
+    {
+      "subset_type": "accent",
+      "subset": "senegal",
+      "cer": 0.1042296072507553,
+      "wer": 0.3274336283185841
+    },
+    {
+      "subset_type": "accent",
+      "subset": "france",
+      "cer": 0.11561092099756272,
+      "wer": 0.3648027291604865
+    },
+    {
+      "subset_type": "accent",
+      "subset": "germany",
+      "cer": 0.13814016172506738,
+      "wer": 0.4291497975708502
+    },
+    {
+      "subset_type": "accent",
+      "subset": "tunisia",
+      "cer": 0.15045871559633028,
+      "wer": 0.40425531914893614
+    },
+    {
+      "subset_type": "accent",
+      "subset": "algeria",
+      "cer": 0.1310592459605027,
+      "wer": 0.44086021505376344
+    },
+    {
+      "subset_type": "accent",
+      "subset": "cameroon",
+      "cer": 0.17333333333333334,
+      "wer": 0.4675324675324675
+    },
+    {
+      "subset_type": "accent",
+      "subset": "new_caledonia",
+      "cer": 0.22140221402214022,
+      "wer": 0.6341463414634146
+    }
+  ]
+}

cv7_transcriptions.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e6c1ed02afd1a2878543a95c6fd0c440042f54f69f9608097f2fac627de51d58
+size 128823777

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 sha256:184295aafb7d6269ba2e4d85970e1c3b8f033f155ea04c0ad1f9a654fb715bb3
+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/tmpc3ob0b51", "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}