jonatasgrosman/exp_w2v2r_fr_vp-100k_age_teens-8_sixties-2_s132

.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_s132
+
+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": 0.10361293102624818,
+      "wer": 0.34501178921271064
+    },
+    {
+      "subset_type": "age",
+      "subset": "",
+      "cer": 0.10131822018944002,
+      "wer": 0.33982808916899365
+    },
+    {
+      "subset_type": "age",
+      "subset": "thirties",
+      "cer": 0.11139567973095331,
+      "wer": 0.3616759088108441
+    },
+    {
+      "subset_type": "age",
+      "subset": "sixties",
+      "cer": 0.10217345726465861,
+      "wer": 0.3381849315068493
+    },
+    {
+      "subset_type": "age",
+      "subset": "seventies",
+      "cer": 0.10464930500079885,
+      "wer": 0.3563218390804598
+    },
+    {
+      "subset_type": "age",
+      "subset": "fifties",
+      "cer": 0.11762264464181954,
+      "wer": 0.3543859649122807
+    },
+    {
+      "subset_type": "age",
+      "subset": "twenties",
+      "cer": 0.11853355367150453,
+      "wer": 0.3806217284478154
+    },
+    {
+      "subset_type": "age",
+      "subset": "eighties",
+      "cer": 0.09714285714285714,
+      "wer": 0.3793103448275862
+    },
+    {
+      "subset_type": "age",
+      "subset": "teens",
+      "cer": 0.1162658944983004,
+      "wer": 0.3710091743119266
+    },
+    {
+      "subset_type": "age",
+      "subset": "fourties",
+      "cer": 0.11487935134292286,
+      "wer": 0.3767107399675249
+    },
+    {
+      "subset_type": "gender",
+      "subset": "",
+      "cer": 0.10132224057231617,
+      "wer": 0.3398125623752495
+    },
+    {
+      "subset_type": "gender",
+      "subset": "male",
+      "cer": 0.11242746981380762,
+      "wer": 0.36519508572971515
+    },
+    {
+      "subset_type": "gender",
+      "subset": "other",
+      "cer": 0.09985802176999527,
+      "wer": 0.32947976878612717
+    },
+    {
+      "subset_type": "gender",
+      "subset": "female",
+      "cer": 0.12163346380791026,
+      "wer": 0.3844256518675123
+    },
+    {
+      "subset_type": "accent",
+      "subset": "",
+      "cer": 0.10232595142513301,
+      "wer": 0.34220966624277543
+    },
+    {
+      "subset_type": "accent",
+      "subset": "united_kingdom",
+      "cer": 0.17343173431734318,
+      "wer": 0.6521739130434783
+    },
+    {
+      "subset_type": "accent",
+      "subset": "st_pierre_et_miquelon",
+      "cer": 0.28222996515679444,
+      "wer": 0.7291666666666666
+    },
+    {
+      "subset_type": "accent",
+      "subset": "martinique",
+      "cer": 0.14919354838709678,
+      "wer": 0.39473684210526316
+    },
+    {
+      "subset_type": "accent",
+      "subset": "romania",
+      "cer": 0.15503875968992248,
+      "wer": 0.6222222222222222
+    },
+    {
+      "subset_type": "accent",
+      "subset": "cote_d_ivoire",
+      "cer": 0.2877059569074778,
+      "wer": 0.6212121212121212
+    },
+    {
+      "subset_type": "accent",
+      "subset": "guadeloupe",
+      "cer": 0.09540636042402827,
+      "wer": 0.3402777777777778
+    },
+    {
+      "subset_type": "accent",
+      "subset": "switzerland",
+      "cer": 0.09446254071661238,
+      "wer": 0.30193905817174516
+    },
+    {
+      "subset_type": "accent",
+      "subset": "united_states",
+      "cer": 0.12836438923395446,
+      "wer": 0.36615384615384616
+    },
+    {
+      "subset_type": "accent",
+      "subset": "canada",
+      "cer": 0.15763025654588733,
+      "wer": 0.4647239263803681
+    },
+    {
+      "subset_type": "accent",
+      "subset": "morocco",
+      "cer": 0.06688963210702341,
+      "wer": 0.25
+    },
+    {
+      "subset_type": "accent",
+      "subset": "belgium",
+      "cer": 0.1237246730852134,
+      "wer": 0.36554621848739494
+    },
+    {
+      "subset_type": "accent",
+      "subset": "luxembourg",
+      "cer": 0.11578947368421053,
+      "wer": 0.38333333333333336
+    },
+    {
+      "subset_type": "accent",
+      "subset": "italy",
+      "cer": 0.2463157894736842,
+      "wer": 0.569620253164557
+    },
+    {
+      "subset_type": "accent",
+      "subset": "monaco",
+      "cer": 0.06557377049180328,
+      "wer": 0.3333333333333333
+    },
+    {
+      "subset_type": "accent",
+      "subset": "senegal",
+      "cer": 0.17832167832167833,
+      "wer": 0.5
+    },
+    {
+      "subset_type": "accent",
+      "subset": "mayotte",
+      "cer": 0.04291845493562232,
+      "wer": 0.18421052631578946
+    },
+    {
+      "subset_type": "accent",
+      "subset": "france",
+      "cer": 0.10367629946472566,
+      "wer": 0.3473137415227743
+    },
+    {
+      "subset_type": "accent",
+      "subset": "germany",
+      "cer": 0.10067114093959731,
+      "wer": 0.3867924528301887
+    },
+    {
+      "subset_type": "accent",
+      "subset": "tunisia",
+      "cer": 0.22297297297297297,
+      "wer": 0.4583333333333333
+    },
+    {
+      "subset_type": "accent",
+      "subset": "gabon",
+      "cer": 0.11481481481481481,
+      "wer": 0.3958333333333333
+    },
+    {
+      "subset_type": "accent",
+      "subset": "algeria",
+      "cer": 0.15736766809728184,
+      "wer": 0.48
+    },
+    {
+      "subset_type": "accent",
+      "subset": "cameroon",
+      "cer": 0.2505050505050505,
+      "wer": 0.5903614457831325
+    }
+  ],
+  "test": [
+    {
+      "subset_type": "all",
+      "subset": "all",
+      "cer": 0.1163951649749132,
+      "wer": 0.36515169760617566
+    },
+    {
+      "subset_type": "age",
+      "subset": "",
+      "cer": 0.11514240837187137,
+      "wer": 0.361257388107253
+    },
+    {
+      "subset_type": "age",
+      "subset": "thirties",
+      "cer": 0.1277604660615093,
+      "wer": 0.403217011995638
+    },
+    {
+      "subset_type": "age",
+      "subset": "seventies",
+      "cer": 0.1100762066045724,
+      "wer": 0.36271186440677966
+    },
+    {
+      "subset_type": "age",
+      "subset": "fifties",
+      "cer": 0.10783308931185945,
+      "wer": 0.3443708609271523
+    },
+    {
+      "subset_type": "age",
+      "subset": "twenties",
+      "cer": 0.13009490623997066,
+      "wer": 0.40577923856477677
+    },
+    {
+      "subset_type": "age",
+      "subset": "fourties",
+      "cer": 0.12377630246427365,
+      "wer": 0.3806366047745358
+    },
+    {
+      "subset_type": "age",
+      "subset": "teens",
+      "cer": 0.12494395456583471,
+      "wer": 0.3904761904761905
+    },
+    {
+      "subset_type": "age",
+      "subset": "sixties",
+      "cer": 0.12206750416225216,
+      "wer": 0.38398544131028206
+    },
+    {
+      "subset_type": "gender",
+      "subset": "",
+      "cer": 0.11506322902851257,
+      "wer": 0.3611499315865339
+    },
+    {
+      "subset_type": "gender",
+      "subset": "male",
+      "cer": 0.1234174638713617,
+      "wer": 0.3864701942397857
+    },
+    {
+      "subset_type": "gender",
+      "subset": "other",
+      "cer": 0.1868563021088769,
+      "wer": 0.46176470588235297
+    },
+    {
+      "subset_type": "gender",
+      "subset": "female",
+      "cer": 0.12641444736541319,
+      "wer": 0.4028512301678547
+    },
+    {
+      "subset_type": "accent",
+      "subset": "",
+      "cer": 0.11608714580454657,
+      "wer": 0.363699186412664
+    },
+    {
+      "subset_type": "accent",
+      "subset": "united_kingdom",
+      "cer": 0.22857142857142856,
+      "wer": 0.6666666666666666
+    },
+    {
+      "subset_type": "accent",
+      "subset": "benin",
+      "cer": 0.09545454545454546,
+      "wer": 0.39473684210526316
+    },
+    {
+      "subset_type": "accent",
+      "subset": "reunion",
+      "cer": 0.10344827586206896,
+      "wer": 0.3333333333333333
+    },
+    {
+      "subset_type": "accent",
+      "subset": "martinique",
+      "cer": 0.09649122807017543,
+      "wer": 0.2894736842105263
+    },
+    {
+      "subset_type": "accent",
+      "subset": "romania",
+      "cer": 0.15267175572519084,
+      "wer": 0.45454545454545453
+    },
+    {
+      "subset_type": "accent",
+      "subset": "cote_d_ivoire",
+      "cer": 0.19134396355353075,
+      "wer": 0.5314685314685315
+    },
+    {
+      "subset_type": "accent",
+      "subset": "other",
+      "cer": 0.04672897196261682,
+      "wer": 0.25
+    },
+    {
+      "subset_type": "accent",
+      "subset": "guadeloupe",
+      "cer": 0.05533596837944664,
+      "wer": 0.21428571428571427
+    },
+    {
+      "subset_type": "accent",
+      "subset": "switzerland",
+      "cer": 0.10116492949110975,
+      "wer": 0.3175182481751825
+    },
+    {
+      "subset_type": "accent",
+      "subset": "united_states",
+      "cer": 0.25264329437952143,
+      "wer": 0.6282051282051282
+    },
+    {
+      "subset_type": "accent",
+      "subset": "canada",
+      "cer": 0.17540609860359077,
+      "wer": 0.48135593220338985
+    },
+    {
+      "subset_type": "accent",
+      "subset": "burundi",
+      "cer": 0.15476190476190477,
+      "wer": 0.5
+    },
+    {
+      "subset_type": "accent",
+      "subset": "morocco",
+      "cer": 0.09523809523809523,
+      "wer": 0.31896551724137934
+    },
+    {
+      "subset_type": "accent",
+      "subset": "belgium",
+      "cer": 0.11102393095331503,
+      "wer": 0.3776223776223776
+    },
+    {
+      "subset_type": "accent",
+      "subset": "monaco",
+      "cer": 0.12158054711246201,
+      "wer": 0.37735849056603776
+    },
+    {
+      "subset_type": "accent",
+      "subset": "senegal",
+      "cer": 0.11782477341389729,
+      "wer": 0.3893805309734513
+    },
+    {
+      "subset_type": "accent",
+      "subset": "france",
+      "cer": 0.11094561658337232,
+      "wer": 0.3602788490062296
+    },
+    {
+      "subset_type": "accent",
+      "subset": "germany",
+      "cer": 0.11657681940700809,
+      "wer": 0.4008097165991903
+    },
+    {
+      "subset_type": "accent",
+      "subset": "tunisia",
+      "cer": 0.11926605504587157,
+      "wer": 0.43617021276595747
+    },
+    {
+      "subset_type": "accent",
+      "subset": "algeria",
+      "cer": 0.12118491921005387,
+      "wer": 0.3978494623655914
+    },
+    {
+      "subset_type": "accent",
+      "subset": "cameroon",
+      "cer": 0.17777777777777778,
+      "wer": 0.45454545454545453
+    },
+    {
+      "subset_type": "accent",
+      "subset": "new_caledonia",
+      "cer": 0.23616236162361623,
+      "wer": 0.5609756097560976
+    }
+  ]
+}

cv7_transcriptions.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6e5d011dd7cdd2ff01b35fec88a3d74c53f1ff05a8348490e251a4a9b070620b
+size 128943278

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:c9eda18b3f2e834e98c811eaa1b48935f359ccdf75184e352cd059103f4dd282
+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/tmprg9d5jnn", "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}