jonatasgrosman/exp_w2v2r_en_xls-r_age_teens-5_sixties-5_s870

.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:
+- en
+license: apache-2.0
+tags:
+- automatic-speech-recognition
+- en
+datasets:
+- mozilla-foundation/common_voice_7_0
+---
+# exp_w2v2r_en_xls-r_age_teens-5_sixties-5_s870
+
+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 (en)](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": 33,
+  "xvector_output_dim": 512
+}

cv7_evaluation.json

@@ -0,0 +1,390 @@
+{
+  "validation": [
+    {
+      "subset_type": "all",
+      "subset": "all",
+      "cer": 0.10515669680934488,
+      "wer": 0.296505376344086
+    },
+    {
+      "subset_type": "age",
+      "subset": "",
+      "cer": 0.1036392394914378,
+      "wer": 0.29292831287084375
+    },
+    {
+      "subset_type": "age",
+      "subset": "sixties",
+      "cer": 0.08175779253960143,
+      "wer": 0.23976023976023977
+    },
+    {
+      "subset_type": "age",
+      "subset": "nineties",
+      "cer": 0.0,
+      "wer": 0.0
+    },
+    {
+      "subset_type": "age",
+      "subset": "fourties",
+      "cer": 0.09394703656998739,
+      "wer": 0.27106227106227104
+    },
+    {
+      "subset_type": "age",
+      "subset": "twenties",
+      "cer": 0.12743700874164401,
+      "wer": 0.3440536589560581
+    },
+    {
+      "subset_type": "age",
+      "subset": "teens",
+      "cer": 0.11164064084364227,
+      "wer": 0.3136390358612581
+    },
+    {
+      "subset_type": "age",
+      "subset": "eighties",
+      "cer": 0.10752688172043011,
+      "wer": 0.28846153846153844
+    },
+    {
+      "subset_type": "age",
+      "subset": "seventies",
+      "cer": 0.07630691399662731,
+      "wer": 0.2583732057416268
+    },
+    {
+      "subset_type": "age",
+      "subset": "thirties",
+      "cer": 0.10977145375558774,
+      "wer": 0.3094498263571559
+    },
+    {
+      "subset_type": "age",
+      "subset": "fifties",
+      "cer": 0.07514840798704803,
+      "wer": 0.25059288537549407
+    },
+    {
+      "subset_type": "gender",
+      "subset": "",
+      "cer": 0.1037034661467019,
+      "wer": 0.29306883400876194
+    },
+    {
+      "subset_type": "gender",
+      "subset": "other",
+      "cer": 0.09643916913946587,
+      "wer": 0.31137724550898205
+    },
+    {
+      "subset_type": "gender",
+      "subset": "male",
+      "cer": 0.11383741202298703,
+      "wer": 0.3159652010345638
+    },
+    {
+      "subset_type": "gender",
+      "subset": "female",
+      "cer": 0.10810279786230745,
+      "wer": 0.3060384263494968
+    },
+    {
+      "subset_type": "accent",
+      "subset": "",
+      "cer": 0.10589747478029794,
+      "wer": 0.29760379696326944
+    },
+    {
+      "subset_type": "accent",
+      "subset": "australia",
+      "cer": 0.06764374295377677,
+      "wer": 0.23076923076923078
+    },
+    {
+      "subset_type": "accent",
+      "subset": "wales",
+      "cer": 0.06779661016949153,
+      "wer": 0.24305555555555555
+    },
+    {
+      "subset_type": "accent",
+      "subset": "singapore",
+      "cer": 0.16590389016018306,
+      "wer": 0.38311688311688313
+    },
+    {
+      "subset_type": "accent",
+      "subset": "england",
+      "cer": 0.09395396770869117,
+      "wer": 0.27805362462760674
+    },
+    {
+      "subset_type": "accent",
+      "subset": "indian",
+      "cer": 0.13688760806916425,
+      "wer": 0.35808779314491884
+    },
+    {
+      "subset_type": "accent",
+      "subset": "scotland",
+      "cer": 0.07473309608540925,
+      "wer": 0.19318181818181818
+    },
+    {
+      "subset_type": "accent",
+      "subset": "southatlandtic",
+      "cer": 0.15384615384615385,
+      "wer": 0.3939393939393939
+    },
+    {
+      "subset_type": "accent",
+      "subset": "ireland",
+      "cer": 0.11396574440052701,
+      "wer": 0.32677165354330706
+    },
+    {
+      "subset_type": "accent",
+      "subset": "us",
+      "cer": 0.08597049464853919,
+      "wer": 0.2610509697789806
+    },
+    {
+      "subset_type": "accent",
+      "subset": "newzealand",
+      "cer": 0.04926108374384237,
+      "wer": 0.15
+    },
+    {
+      "subset_type": "accent",
+      "subset": "other",
+      "cer": 0.08921933085501858,
+      "wer": 0.3
+    },
+    {
+      "subset_type": "accent",
+      "subset": "african",
+      "cer": 0.13323572474377746,
+      "wer": 0.37317784256559766
+    },
+    {
+      "subset_type": "accent",
+      "subset": "malaysia",
+      "cer": 0.11538461538461539,
+      "wer": 0.3472222222222222
+    },
+    {
+      "subset_type": "accent",
+      "subset": "hongkong",
+      "cer": 0.12338858195211787,
+      "wer": 0.358695652173913
+    },
+    {
+      "subset_type": "accent",
+      "subset": "bermuda",
+      "cer": 0.06172839506172839,
+      "wer": 0.125
+    },
+    {
+      "subset_type": "accent",
+      "subset": "canada",
+      "cer": 0.07087576374745418,
+      "wer": 0.23261390887290168
+    },
+    {
+      "subset_type": "accent",
+      "subset": "philippines",
+      "cer": 0.15764425936942297,
+      "wer": 0.42560553633217996
+    }
+  ],
+  "test": [
+    {
+      "subset_type": "all",
+      "subset": "all",
+      "cer": 0.1366993998591328,
+      "wer": 0.34435362584188844
+    },
+    {
+      "subset_type": "age",
+      "subset": "",
+      "cer": 0.13553293083106716,
+      "wer": 0.3414396665170659
+    },
+    {
+      "subset_type": "age",
+      "subset": "sixties",
+      "cer": 0.14234161988773056,
+      "wer": 0.27973568281938327
+    },
+    {
+      "subset_type": "age",
+      "subset": "fourties",
+      "cer": 0.1348998394049531,
+      "wer": 0.33035714285714285
+    },
+    {
+      "subset_type": "age",
+      "subset": "twenties",
+      "cer": 0.14654088050314465,
+      "wer": 0.37248737734590837
+    },
+    {
+      "subset_type": "age",
+      "subset": "teens",
+      "cer": 0.1454025385853048,
+      "wer": 0.37814840027229407
+    },
+    {
+      "subset_type": "age",
+      "subset": "eighties",
+      "cer": 0.14857142857142858,
+      "wer": 0.48
+    },
+    {
+      "subset_type": "age",
+      "subset": "seventies",
+      "cer": 0.10548523206751055,
+      "wer": 0.2651006711409396
+    },
+    {
+      "subset_type": "age",
+      "subset": "thirties",
+      "cer": 0.13256174167455467,
+      "wer": 0.3524390243902439
+    },
+    {
+      "subset_type": "age",
+      "subset": "fifties",
+      "cer": 0.18349414197105446,
+      "wer": 0.3780369290573372
+    },
+    {
+      "subset_type": "gender",
+      "subset": "",
+      "cer": 0.13546165986128694,
+      "wer": 0.3412894542744824
+    },
+    {
+      "subset_type": "gender",
+      "subset": "other",
+      "cer": 0.1774081774081774,
+      "wer": 0.39299610894941633
+    },
+    {
+      "subset_type": "gender",
+      "subset": "male",
+      "cer": 0.1448275182143526,
+      "wer": 0.365791987413609
+    },
+    {
+      "subset_type": "gender",
+      "subset": "female",
+      "cer": 0.13820798998643996,
+      "wer": 0.3470115967885816
+    },
+    {
+      "subset_type": "accent",
+      "subset": "",
+      "cer": 0.13717309584058504,
+      "wer": 0.3450693075007557
+    },
+    {
+      "subset_type": "accent",
+      "subset": "australia",
+      "cer": 0.11509635974304068,
+      "wer": 0.3375796178343949
+    },
+    {
+      "subset_type": "accent",
+      "subset": "singapore",
+      "cer": 0.13368983957219252,
+      "wer": 0.3333333333333333
+    },
+    {
+      "subset_type": "accent",
+      "subset": "wales",
+      "cer": 0.18493150684931506,
+      "wer": 0.375
+    },
+    {
+      "subset_type": "accent",
+      "subset": "indian",
+      "cer": 0.15631868131868132,
+      "wer": 0.39836632108074144
+    },
+    {
+      "subset_type": "accent",
+      "subset": "scotland",
+      "cer": 0.11921458625525946,
+      "wer": 0.3049645390070922
+    },
+    {
+      "subset_type": "accent",
+      "subset": "england",
+      "cer": 0.11918173732582271,
+      "wer": 0.32843414358394574
+    },
+    {
+      "subset_type": "accent",
+      "subset": "ireland",
+      "cer": 0.09941944847605225,
+      "wer": 0.270042194092827
+    },
+    {
+      "subset_type": "accent",
+      "subset": "us",
+      "cer": 0.11859043704024232,
+      "wer": 0.3087268836924958
+    },
+    {
+      "subset_type": "accent",
+      "subset": "newzealand",
+      "cer": 0.08221024258760108,
+      "wer": 0.2647058823529412
+    },
+    {
+      "subset_type": "accent",
+      "subset": "other",
+      "cer": 0.1934477379095164,
+      "wer": 0.416988416988417
+    },
+    {
+      "subset_type": "accent",
+      "subset": "african",
+      "cer": 0.1624874623871615,
+      "wer": 0.3930635838150289
+    },
+    {
+      "subset_type": "accent",
+      "subset": "malaysia",
+      "cer": 0.1019036954087346,
+      "wer": 0.3184713375796178
+    },
+    {
+      "subset_type": "accent",
+      "subset": "hongkong",
+      "cer": 0.2514450867052023,
+      "wer": 0.5666666666666667
+    },
+    {
+      "subset_type": "accent",
+      "subset": "bermuda",
+      "cer": 0.15730337078651685,
+      "wer": 0.35714285714285715
+    },
+    {
+      "subset_type": "accent",
+      "subset": "canada",
+      "cer": 0.16708043694141012,
+      "wer": 0.3160112359550562
+    },
+    {
+      "subset_type": "accent",
+      "subset": "philippines",
+      "cer": 0.10367892976588629,
+      "wer": 0.29292929292929293
+    }
+  ]
+}

cv7_transcriptions.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:613ad517a93d156d2905ff14481a6bc5697f9fb01f56710e514e4c8c59a5d73b
+size 123301687

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 sha256:9971666d2603b86e8bb6e44bc788f3d920cb74a980032456f661dbd97fe29438
+size 1262034033

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/tmp1olpxwgu", "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}