jonatasgrosman/exp_w2v2r_fr_vp-100k_age_teens-8_sixties-2_s607

.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-8_sixties-2_s607
+
+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": 0.10397803436429034,
+      "wer": 0.34795110511888255
+    },
+    {
+      "subset_type": "age",
+      "subset": "",
+      "cer": 0.1019477908791108,
+      "wer": 0.34328346567243345
+    },
+    {
+      "subset_type": "age",
+      "subset": "thirties",
+      "cer": 0.11362048894062864,
+      "wer": 0.3701478743068392
+    },
+    {
+      "subset_type": "age",
+      "subset": "sixties",
+      "cer": 0.09770074778111677,
+      "wer": 0.3309075342465753
+    },
+    {
+      "subset_type": "age",
+      "subset": "seventies",
+      "cer": 0.11040102252756032,
+      "wer": 0.3716475095785441
+    },
+    {
+      "subset_type": "age",
+      "subset": "fifties",
+      "cer": 0.1213580144434299,
+      "wer": 0.3649122807017544
+    },
+    {
+      "subset_type": "age",
+      "subset": "twenties",
+      "cer": 0.11431240120706998,
+      "wer": 0.37538724495246234
+    },
+    {
+      "subset_type": "age",
+      "subset": "eighties",
+      "cer": 0.10285714285714286,
+      "wer": 0.4482758620689655
+    },
+    {
+      "subset_type": "age",
+      "subset": "teens",
+      "cer": 0.11066347727558856,
+      "wer": 0.3563302752293578
+    },
+    {
+      "subset_type": "age",
+      "subset": "fourties",
+      "cer": 0.11788094959653841,
+      "wer": 0.38343771746694505
+    },
+    {
+      "subset_type": "gender",
+      "subset": "",
+      "cer": 0.10195027899896361,
+      "wer": 0.3432821856287425
+    },
+    {
+      "subset_type": "gender",
+      "subset": "male",
+      "cer": 0.11169269925916948,
+      "wer": 0.3665451599837991
+    },
+    {
+      "subset_type": "gender",
+      "subset": "other",
+      "cer": 0.10033128253667771,
+      "wer": 0.33236994219653176
+    },
+    {
+      "subset_type": "gender",
+      "subset": "female",
+      "cer": 0.12028977040369224,
+      "wer": 0.38160676532769555
+    },
+    {
+      "subset_type": "accent",
+      "subset": "",
+      "cer": 0.10288537185041881,
+      "wer": 0.34546119110894513
+    },
+    {
+      "subset_type": "accent",
+      "subset": "united_kingdom",
+      "cer": 0.1881918819188192,
+      "wer": 0.6956521739130435
+    },
+    {
+      "subset_type": "accent",
+      "subset": "st_pierre_et_miquelon",
+      "cer": 0.2578397212543554,
+      "wer": 0.6875
+    },
+    {
+      "subset_type": "accent",
+      "subset": "martinique",
+      "cer": 0.16129032258064516,
+      "wer": 0.39473684210526316
+    },
+    {
+      "subset_type": "accent",
+      "subset": "romania",
+      "cer": 0.14728682170542637,
+      "wer": 0.5333333333333333
+    },
+    {
+      "subset_type": "accent",
+      "subset": "cote_d_ivoire",
+      "cer": 0.302915082382763,
+      "wer": 0.6515151515151515
+    },
+    {
+      "subset_type": "accent",
+      "subset": "guadeloupe",
+      "cer": 0.10836277974087162,
+      "wer": 0.3611111111111111
+    },
+    {
+      "subset_type": "accent",
+      "subset": "switzerland",
+      "cer": 0.09120521172638436,
+      "wer": 0.3047091412742382
+    },
+    {
+      "subset_type": "accent",
+      "subset": "united_states",
+      "cer": 0.12008281573498965,
+      "wer": 0.3446153846153846
+    },
+    {
+      "subset_type": "accent",
+      "subset": "canada",
+      "cer": 0.14824120603015076,
+      "wer": 0.44785276073619634
+    },
+    {
+      "subset_type": "accent",
+      "subset": "morocco",
+      "cer": 0.03678929765886288,
+      "wer": 0.1346153846153846
+    },
+    {
+      "subset_type": "accent",
+      "subset": "belgium",
+      "cer": 0.12444316712171288,
+      "wer": 0.3798319327731092
+    },
+    {
+      "subset_type": "accent",
+      "subset": "luxembourg",
+      "cer": 0.12631578947368421,
+      "wer": 0.43333333333333335
+    },
+    {
+      "subset_type": "accent",
+      "subset": "italy",
+      "cer": 0.24210526315789474,
+      "wer": 0.569620253164557
+    },
+    {
+      "subset_type": "accent",
+      "subset": "monaco",
+      "cer": 0.07103825136612021,
+      "wer": 0.3333333333333333
+    },
+    {
+      "subset_type": "accent",
+      "subset": "senegal",
+      "cer": 0.17132867132867133,
+      "wer": 0.5
+    },
+    {
+      "subset_type": "accent",
+      "subset": "mayotte",
+      "cer": 0.034334763948497854,
+      "wer": 0.21052631578947367
+    },
+    {
+      "subset_type": "accent",
+      "subset": "france",
+      "cer": 0.10353184761935334,
+      "wer": 0.3494781395796258
+    },
+    {
+      "subset_type": "accent",
+      "subset": "germany",
+      "cer": 0.09060402684563758,
+      "wer": 0.39622641509433965
+    },
+    {
+      "subset_type": "accent",
+      "subset": "tunisia",
+      "cer": 0.24324324324324326,
+      "wer": 0.5
+    },
+    {
+      "subset_type": "accent",
+      "subset": "gabon",
+      "cer": 0.11851851851851852,
+      "wer": 0.3541666666666667
+    },
+    {
+      "subset_type": "accent",
+      "subset": "algeria",
+      "cer": 0.1473533619456366,
+      "wer": 0.464
+    },
+    {
+      "subset_type": "accent",
+      "subset": "cameroon",
+      "cer": 0.2545454545454545,
+      "wer": 0.5783132530120482
+    }
+  ],
+  "test": [
+    {
+      "subset_type": "all",
+      "subset": "all",
+      "cer": 0.11683731584942447,
+      "wer": 0.36687560606156633
+    },
+    {
+      "subset_type": "age",
+      "subset": "",
+      "cer": 0.11541837474044572,
+      "wer": 0.36305177351194184
+    },
+    {
+      "subset_type": "age",
+      "subset": "thirties",
+      "cer": 0.13001851600957412,
+      "wer": 0.3999454743729553
+    },
+    {
+      "subset_type": "age",
+      "subset": "seventies",
+      "cer": 0.1143099068585944,
+      "wer": 0.36271186440677966
+    },
+    {
+      "subset_type": "age",
+      "subset": "fifties",
+      "cer": 0.11251830161054173,
+      "wer": 0.354083885209713
+    },
+    {
+      "subset_type": "age",
+      "subset": "twenties",
+      "cer": 0.13202054009444775,
+      "wer": 0.4129005751848809
+    },
+    {
+      "subset_type": "age",
+      "subset": "fourties",
+      "cer": 0.12265106335096208,
+      "wer": 0.3789787798408488
+    },
+    {
+      "subset_type": "age",
+      "subset": "teens",
+      "cer": 0.12696159019578537,
+      "wer": 0.39183673469387753
+    },
+    {
+      "subset_type": "age",
+      "subset": "sixties",
+      "cer": 0.1200242167398214,
+      "wer": 0.3689717925386715
+    },
+    {
+      "subset_type": "gender",
+      "subset": "",
+      "cer": 0.11532799906723215,
+      "wer": 0.36291331611411515
+    },
+    {
+      "subset_type": "gender",
+      "subset": "male",
+      "cer": 0.12513739059637696,
+      "wer": 0.38720087681909515
+    },
+    {
+      "subset_type": "gender",
+      "subset": "other",
+      "cer": 0.19077979401667483,
+      "wer": 0.47058823529411764
+    },
+    {
+      "subset_type": "gender",
+      "subset": "female",
+      "cer": 0.12736693717377223,
+      "wer": 0.40653023683605427
+    },
+    {
+      "subset_type": "accent",
+      "subset": "",
+      "cer": 0.11635687380515228,
+      "wer": 0.3654659960389268
+    },
+    {
+      "subset_type": "accent",
+      "subset": "united_kingdom",
+      "cer": 0.2571428571428571,
+      "wer": 0.7777777777777778
+    },
+    {
+      "subset_type": "accent",
+      "subset": "benin",
+      "cer": 0.11818181818181818,
+      "wer": 0.39473684210526316
+    },
+    {
+      "subset_type": "accent",
+      "subset": "reunion",
+      "cer": 0.109717868338558,
+      "wer": 0.3333333333333333
+    },
+    {
+      "subset_type": "accent",
+      "subset": "martinique",
+      "cer": 0.10964912280701754,
+      "wer": 0.3157894736842105
+    },
+    {
+      "subset_type": "accent",
+      "subset": "romania",
+      "cer": 0.15267175572519084,
+      "wer": 0.36363636363636365
+    },
+    {
+      "subset_type": "accent",
+      "subset": "cote_d_ivoire",
+      "cer": 0.2152619589977221,
+      "wer": 0.5524475524475524
+    },
+    {
+      "subset_type": "accent",
+      "subset": "other",
+      "cer": 0.037383177570093455,
+      "wer": 0.2
+    },
+    {
+      "subset_type": "accent",
+      "subset": "guadeloupe",
+      "cer": 0.05138339920948617,
+      "wer": 0.23809523809523808
+    },
+    {
+      "subset_type": "accent",
+      "subset": "switzerland",
+      "cer": 0.09687308399754752,
+      "wer": 0.3284671532846715
+    },
+    {
+      "subset_type": "accent",
+      "subset": "united_states",
+      "cer": 0.23817473567056205,
+      "wer": 0.6121794871794872
+    },
+    {
+      "subset_type": "accent",
+      "subset": "canada",
+      "cer": 0.17540609860359077,
+      "wer": 0.4754237288135593
+    },
+    {
+      "subset_type": "accent",
+      "subset": "burundi",
+      "cer": 0.15476190476190477,
+      "wer": 0.5
+    },
+    {
+      "subset_type": "accent",
+      "subset": "morocco",
+      "cer": 0.09103641456582633,
+      "wer": 0.3103448275862069
+    },
+    {
+      "subset_type": "accent",
+      "subset": "belgium",
+      "cer": 0.1116123970184386,
+      "wer": 0.3706293706293706
+    },
+    {
+      "subset_type": "accent",
+      "subset": "monaco",
+      "cer": 0.1276595744680851,
+      "wer": 0.39622641509433965
+    },
+    {
+      "subset_type": "accent",
+      "subset": "senegal",
+      "cer": 0.09969788519637462,
+      "wer": 0.36283185840707965
+    },
+    {
+      "subset_type": "accent",
+      "subset": "france",
+      "cer": 0.11371525640710013,
+      "wer": 0.36287451794719666
+    },
+    {
+      "subset_type": "accent",
+      "subset": "germany",
+      "cer": 0.12129380053908356,
+      "wer": 0.41700404858299595
+    },
+    {
+      "subset_type": "accent",
+      "subset": "tunisia",
+      "cer": 0.12110091743119267,
+      "wer": 0.39361702127659576
+    },
+    {
+      "subset_type": "accent",
+      "subset": "algeria",
+      "cer": 0.11579892280071813,
+      "wer": 0.44086021505376344
+    },
+    {
+      "subset_type": "accent",
+      "subset": "cameroon",
+      "cer": 0.15333333333333332,
+      "wer": 0.42857142857142855
+    },
+    {
+      "subset_type": "accent",
+      "subset": "new_caledonia",
+      "cer": 0.19557195571955718,
+      "wer": 0.5121951219512195
+    }
+  ]
+}

cv7_transcriptions.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:446b73aa62e0eb924fedcd001f820d3c79a0a66ff137e370f88716672f03d48a
+size 129151627

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 sha256:1faa4c20b3fc33725521ae17a6f9ca317c875ebf9ed68d84ba900dcdd24f1d73
+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/tmpoem2ak4f", "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}