jonatasgrosman/exp_w2v2r_fr_vp-100k_age_teens-0_sixties-10_s151

.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_s151
+
+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": 0.10543101828806865,
+      "wer": 0.34519070409395586
+    },
+    {
+      "subset_type": "age",
+      "subset": "",
+      "cer": 0.10288115851348924,
+      "wer": 0.33937569224529274
+    },
+    {
+      "subset_type": "age",
+      "subset": "seventies",
+      "cer": 0.10960217287106566,
+      "wer": 0.36590038314176243
+    },
+    {
+      "subset_type": "age",
+      "subset": "sixties",
+      "cer": 0.10007687469424838,
+      "wer": 0.3356164383561644
+    },
+    {
+      "subset_type": "age",
+      "subset": "thirties",
+      "cer": 0.1165437847626439,
+      "wer": 0.3659889094269871
+    },
+    {
+      "subset_type": "age",
+      "subset": "eighties",
+      "cer": 0.08,
+      "wer": 0.3448275862068966
+    },
+    {
+      "subset_type": "age",
+      "subset": "twenties",
+      "cer": 0.11982684293720362,
+      "wer": 0.38030124986646724
+    },
+    {
+      "subset_type": "age",
+      "subset": "fourties",
+      "cer": 0.11659455034498889,
+      "wer": 0.37531895151936906
+    },
+    {
+      "subset_type": "age",
+      "subset": "fifties",
+      "cer": 0.12243712127500622,
+      "wer": 0.36641604010025064
+    },
+    {
+      "subset_type": "age",
+      "subset": "teens",
+      "cer": 0.12463804607830795,
+      "wer": 0.3856880733944954
+    },
+    {
+      "subset_type": "gender",
+      "subset": "",
+      "cer": 0.10287682951728915,
+      "wer": 0.33935254491017963
+    },
+    {
+      "subset_type": "gender",
+      "subset": "other",
+      "cer": 0.11547562707051585,
+      "wer": 0.36127167630057805
+    },
+    {
+      "subset_type": "gender",
+      "subset": "female",
+      "cer": 0.12522638312788456,
+      "wer": 0.3870683579985906
+    },
+    {
+      "subset_type": "gender",
+      "subset": "male",
+      "cer": 0.11510900359052828,
+      "wer": 0.36794023671301923
+    },
+    {
+      "subset_type": "accent",
+      "subset": "",
+      "cer": 0.10396759149933514,
+      "wer": 0.34187461450012563
+    },
+    {
+      "subset_type": "accent",
+      "subset": "italy",
+      "cer": 0.23789473684210527,
+      "wer": 0.569620253164557
+    },
+    {
+      "subset_type": "accent",
+      "subset": "united_states",
+      "cer": 0.14855072463768115,
+      "wer": 0.39692307692307693
+    },
+    {
+      "subset_type": "accent",
+      "subset": "germany",
+      "cer": 0.1040268456375839,
+      "wer": 0.4056603773584906
+    },
+    {
+      "subset_type": "accent",
+      "subset": "gabon",
+      "cer": 0.1,
+      "wer": 0.3541666666666667
+    },
+    {
+      "subset_type": "accent",
+      "subset": "martinique",
+      "cer": 0.16532258064516128,
+      "wer": 0.4473684210526316
+    },
+    {
+      "subset_type": "accent",
+      "subset": "belgium",
+      "cer": 0.11941370886621641,
+      "wer": 0.36554621848739494
+    },
+    {
+      "subset_type": "accent",
+      "subset": "monaco",
+      "cer": 0.05191256830601093,
+      "wer": 0.2857142857142857
+    },
+    {
+      "subset_type": "accent",
+      "subset": "algeria",
+      "cer": 0.1759656652360515,
+      "wer": 0.512
+    },
+    {
+      "subset_type": "accent",
+      "subset": "st_pierre_et_miquelon",
+      "cer": 0.2857142857142857,
+      "wer": 0.6666666666666666
+    },
+    {
+      "subset_type": "accent",
+      "subset": "france",
+      "cer": 0.10619618165622066,
+      "wer": 0.3493819441104324
+    },
+    {
+      "subset_type": "accent",
+      "subset": "guadeloupe",
+      "cer": 0.10954063604240283,
+      "wer": 0.3680555555555556
+    },
+    {
+      "subset_type": "accent",
+      "subset": "canada",
+      "cer": 0.16754826765405978,
+      "wer": 0.4754601226993865
+    },
+    {
+      "subset_type": "accent",
+      "subset": "cameroon",
+      "cer": 0.23434343434343435,
+      "wer": 0.5542168674698795
+    },
+    {
+      "subset_type": "accent",
+      "subset": "luxembourg",
+      "cer": 0.13157894736842105,
+      "wer": 0.4
+    },
+    {
+      "subset_type": "accent",
+      "subset": "tunisia",
+      "cer": 0.25675675675675674,
+      "wer": 0.5416666666666666
+    },
+    {
+      "subset_type": "accent",
+      "subset": "switzerland",
+      "cer": 0.08887854816193579,
+      "wer": 0.29362880886426596
+    },
+    {
+      "subset_type": "accent",
+      "subset": "romania",
+      "cer": 0.17829457364341086,
+      "wer": 0.5777777777777777
+    },
+    {
+      "subset_type": "accent",
+      "subset": "united_kingdom",
+      "cer": 0.15498154981549817,
+      "wer": 0.5434782608695652
+    },
+    {
+      "subset_type": "accent",
+      "subset": "mayotte",
+      "cer": 0.03862660944206009,
+      "wer": 0.18421052631578946
+    },
+    {
+      "subset_type": "accent",
+      "subset": "senegal",
+      "cer": 0.15034965034965034,
+      "wer": 0.48
+    },
+    {
+      "subset_type": "accent",
+      "subset": "cote_d_ivoire",
+      "cer": 0.3041825095057034,
+      "wer": 0.7121212121212122
+    },
+    {
+      "subset_type": "accent",
+      "subset": "morocco",
+      "cer": 0.05016722408026756,
+      "wer": 0.21153846153846154
+    }
+  ],
+  "test": [
+    {
+      "subset_type": "all",
+      "subset": "all",
+      "cer": 0.11977024998368253,
+      "wer": 0.36765516760573197
+    },
+    {
+      "subset_type": "age",
+      "subset": "",
+      "cer": 0.11840537635977261,
+      "wer": 0.36363769527673817
+    },
+    {
+      "subset_type": "age",
+      "subset": "seventies",
+      "cer": 0.10866497318656505,
+      "wer": 0.3440677966101695
+    },
+    {
+      "subset_type": "age",
+      "subset": "sixties",
+      "cer": 0.11608899651884365,
+      "wer": 0.3630573248407643
+    },
+    {
+      "subset_type": "age",
+      "subset": "thirties",
+      "cer": 0.13141850697737434,
+      "wer": 0.4010359869138495
+    },
+    {
+      "subset_type": "age",
+      "subset": "twenties",
+      "cer": 0.13612397414148825,
+      "wer": 0.4130375239660367
+    },
+    {
+      "subset_type": "age",
+      "subset": "fourties",
+      "cer": 0.1239450883312704,
+      "wer": 0.378315649867374
+    },
+    {
+      "subset_type": "age",
+      "subset": "fifties",
+      "cer": 0.11456808199121522,
+      "wer": 0.3615894039735099
+    },
+    {
+      "subset_type": "age",
+      "subset": "teens",
+      "cer": 0.13742340457330743,
+      "wer": 0.41315192743764173
+    },
+    {
+      "subset_type": "gender",
+      "subset": "female",
+      "cer": 0.13216748580790186,
+      "wer": 0.408829616003679
+    },
+    {
+      "subset_type": "gender",
+      "subset": "",
+      "cer": 0.11830484019063443,
+      "wer": 0.36346940418090423
+    },
+    {
+      "subset_type": "gender",
+      "subset": "other",
+      "cer": 0.19911721432074547,
+      "wer": 0.5205882352941177
+    },
+    {
+      "subset_type": "gender",
+      "subset": "male",
+      "cer": 0.12709139018929372,
+      "wer": 0.38841868111794436
+    },
+    {
+      "subset_type": "accent",
+      "subset": "",
+      "cer": 0.11930613654861728,
+      "wer": 0.36632090392260236
+    },
+    {
+      "subset_type": "accent",
+      "subset": "reunion",
+      "cer": 0.10344827586206896,
+      "wer": 0.35185185185185186
+    },
+    {
+      "subset_type": "accent",
+      "subset": "united_states",
+      "cer": 0.27712854757929883,
+      "wer": 0.6474358974358975
+    },
+    {
+      "subset_type": "accent",
+      "subset": "other",
+      "cer": 0.04672897196261682,
+      "wer": 0.2
+    },
+    {
+      "subset_type": "accent",
+      "subset": "germany",
+      "cer": 0.1421832884097035,
+      "wer": 0.4534412955465587
+    },
+    {
+      "subset_type": "accent",
+      "subset": "martinique",
+      "cer": 0.11842105263157894,
+      "wer": 0.2894736842105263
+    },
+    {
+      "subset_type": "accent",
+      "subset": "belgium",
+      "cer": 0.10867006669282071,
+      "wer": 0.36713286713286714
+    },
+    {
+      "subset_type": "accent",
+      "subset": "algeria",
+      "cer": 0.12387791741472172,
+      "wer": 0.3978494623655914
+    },
+    {
+      "subset_type": "accent",
+      "subset": "monaco",
+      "cer": 0.12462006079027356,
+      "wer": 0.37735849056603776
+    },
+    {
+      "subset_type": "accent",
+      "subset": "burundi",
+      "cer": 0.17857142857142858,
+      "wer": 0.6428571428571429
+    },
+    {
+      "subset_type": "accent",
+      "subset": "france",
+      "cer": 0.11500775499150644,
+      "wer": 0.36013052506674575
+    },
+    {
+      "subset_type": "accent",
+      "subset": "guadeloupe",
+      "cer": 0.03162055335968379,
+      "wer": 0.19047619047619047
+    },
+    {
+      "subset_type": "accent",
+      "subset": "canada",
+      "cer": 0.1811057281276717,
+      "wer": 0.49322033898305084
+    },
+    {
+      "subset_type": "accent",
+      "subset": "cameroon",
+      "cer": 0.19333333333333333,
+      "wer": 0.4935064935064935
+    },
+    {
+      "subset_type": "accent",
+      "subset": "new_caledonia",
+      "cer": 0.2177121771217712,
+      "wer": 0.5121951219512195
+    },
+    {
+      "subset_type": "accent",
+      "subset": "tunisia",
+      "cer": 0.13394495412844037,
+      "wer": 0.425531914893617
+    },
+    {
+      "subset_type": "accent",
+      "subset": "switzerland",
+      "cer": 0.12262415695892091,
+      "wer": 0.3795620437956204
+    },
+    {
+      "subset_type": "accent",
+      "subset": "benin",
+      "cer": 0.09545454545454546,
+      "wer": 0.3684210526315789
+    },
+    {
+      "subset_type": "accent",
+      "subset": "romania",
+      "cer": 0.15267175572519084,
+      "wer": 0.36363636363636365
+    },
+    {
+      "subset_type": "accent",
+      "subset": "united_kingdom",
+      "cer": 0.22857142857142856,
+      "wer": 0.4444444444444444
+    },
+    {
+      "subset_type": "accent",
+      "subset": "senegal",
+      "cer": 0.0891238670694864,
+      "wer": 0.2920353982300885
+    },
+    {
+      "subset_type": "accent",
+      "subset": "cote_d_ivoire",
+      "cer": 0.2061503416856492,
+      "wer": 0.5384615384615384
+    },
+    {
+      "subset_type": "accent",
+      "subset": "morocco",
+      "cer": 0.09663865546218488,
+      "wer": 0.33620689655172414
+    }
+  ]
+}

cv7_transcriptions.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3dcbe475f2b348113282c027500cc09ff46b983716083d7351f461ceb4cc49db
+size 128872228

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:e07c4357bdfffa32512a66b66af34109fa678b438932dab46d99bee91a260562
+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/tmpbqe863kd", "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}