jonatasgrosman/exp_w2v2r_en_xls-r_age_teens-0_sixties-10_s732

.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-0_sixties-10_s732
+
+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": 0.11930850161025926,
+      "wer": 0.32456864216054016
+    },
+    {
+      "subset_type": "age",
+      "subset": "",
+      "cer": 0.11709989724590812,
+      "wer": 0.3197895286785834
+    },
+    {
+      "subset_type": "age",
+      "subset": "sixties",
+      "cer": 0.08226877874297393,
+      "wer": 0.24075924075924077
+    },
+    {
+      "subset_type": "age",
+      "subset": "nineties",
+      "cer": 0.07692307692307693,
+      "wer": 0.2857142857142857
+    },
+    {
+      "subset_type": "age",
+      "subset": "fourties",
+      "cer": 0.10680958385876418,
+      "wer": 0.30256410256410254
+    },
+    {
+      "subset_type": "age",
+      "subset": "twenties",
+      "cer": 0.14735914199662087,
+      "wer": 0.3814601427465818
+    },
+    {
+      "subset_type": "age",
+      "subset": "teens",
+      "cer": 0.13567227742851348,
+      "wer": 0.36037624926513817
+    },
+    {
+      "subset_type": "age",
+      "subset": "eighties",
+      "cer": 0.11469534050179211,
+      "wer": 0.3076923076923077
+    },
+    {
+      "subset_type": "age",
+      "subset": "seventies",
+      "cer": 0.08473861720067453,
+      "wer": 0.2631578947368421
+    },
+    {
+      "subset_type": "age",
+      "subset": "thirties",
+      "cer": 0.12655039979852672,
+      "wer": 0.3412538841162493
+    },
+    {
+      "subset_type": "age",
+      "subset": "fifties",
+      "cer": 0.0868861305990286,
+      "wer": 0.27430830039525694
+    },
+    {
+      "subset_type": "gender",
+      "subset": "",
+      "cer": 0.11717038253328886,
+      "wer": 0.3199143213248849
+    },
+    {
+      "subset_type": "gender",
+      "subset": "other",
+      "cer": 0.12314540059347182,
+      "wer": 0.37425149700598803
+    },
+    {
+      "subset_type": "gender",
+      "subset": "male",
+      "cer": 0.13223187189255506,
+      "wer": 0.3512814483893722
+    },
+    {
+      "subset_type": "gender",
+      "subset": "female",
+      "cer": 0.12158126375353662,
+      "wer": 0.3334858188472095
+    },
+    {
+      "subset_type": "accent",
+      "subset": "",
+      "cer": 0.11995936926072755,
+      "wer": 0.3254859958978236
+    },
+    {
+      "subset_type": "accent",
+      "subset": "australia",
+      "cer": 0.07666290868094701,
+      "wer": 0.248868778280543
+    },
+    {
+      "subset_type": "accent",
+      "subset": "wales",
+      "cer": 0.07457627118644068,
+      "wer": 0.2708333333333333
+    },
+    {
+      "subset_type": "accent",
+      "subset": "singapore",
+      "cer": 0.17963386727688788,
+      "wer": 0.3961038961038961
+    },
+    {
+      "subset_type": "accent",
+      "subset": "england",
+      "cer": 0.10623497080041223,
+      "wer": 0.3138033763654419
+    },
+    {
+      "subset_type": "accent",
+      "subset": "indian",
+      "cer": 0.16730135858377934,
+      "wer": 0.41581479254359593
+    },
+    {
+      "subset_type": "accent",
+      "subset": "scotland",
+      "cer": 0.0800711743772242,
+      "wer": 0.25
+    },
+    {
+      "subset_type": "accent",
+      "subset": "southatlandtic",
+      "cer": 0.17032967032967034,
+      "wer": 0.48484848484848486
+    },
+    {
+      "subset_type": "accent",
+      "subset": "ireland",
+      "cer": 0.12582345191040845,
+      "wer": 0.3700787401574803
+    },
+    {
+      "subset_type": "accent",
+      "subset": "us",
+      "cer": 0.09667341625687012,
+      "wer": 0.28258908434821833
+    },
+    {
+      "subset_type": "accent",
+      "subset": "newzealand",
+      "cer": 0.054187192118226604,
+      "wer": 0.15
+    },
+    {
+      "subset_type": "accent",
+      "subset": "other",
+      "cer": 0.1171003717472119,
+      "wer": 0.38
+    },
+    {
+      "subset_type": "accent",
+      "subset": "african",
+      "cer": 0.14202049780380674,
+      "wer": 0.36151603498542273
+    },
+    {
+      "subset_type": "accent",
+      "subset": "malaysia",
+      "cer": 0.14102564102564102,
+      "wer": 0.3194444444444444
+    },
+    {
+      "subset_type": "accent",
+      "subset": "hongkong",
+      "cer": 0.11786372007366483,
+      "wer": 0.3695652173913043
+    },
+    {
+      "subset_type": "accent",
+      "subset": "bermuda",
+      "cer": 0.08641975308641975,
+      "wer": 0.1875
+    },
+    {
+      "subset_type": "accent",
+      "subset": "canada",
+      "cer": 0.08391038696537678,
+      "wer": 0.2517985611510791
+    },
+    {
+      "subset_type": "accent",
+      "subset": "philippines",
+      "cer": 0.17430101130279596,
+      "wer": 0.4186851211072664
+    }
+  ],
+  "test": [
+    {
+      "subset_type": "all",
+      "subset": "all",
+      "cer": 0.1540656487749792,
+      "wer": 0.37408618322108156
+    },
+    {
+      "subset_type": "age",
+      "subset": "",
+      "cer": 0.15227564933698454,
+      "wer": 0.3695925027194377
+    },
+    {
+      "subset_type": "age",
+      "subset": "sixties",
+      "cer": 0.14514835605453089,
+      "wer": 0.28854625550660795
+    },
+    {
+      "subset_type": "age",
+      "subset": "fourties",
+      "cer": 0.14571887414419746,
+      "wer": 0.34962406015037595
+    },
+    {
+      "subset_type": "age",
+      "subset": "twenties",
+      "cer": 0.17153260509764978,
+      "wer": 0.420120034295513
+    },
+    {
+      "subset_type": "age",
+      "subset": "teens",
+      "cer": 0.1706096180203802,
+      "wer": 0.42239618788291355
+    },
+    {
+      "subset_type": "age",
+      "subset": "eighties",
+      "cer": 0.13714285714285715,
+      "wer": 0.52
+    },
+    {
+      "subset_type": "age",
+      "subset": "seventies",
+      "cer": 0.11693791440626884,
+      "wer": 0.30201342281879195
+    },
+    {
+      "subset_type": "age",
+      "subset": "thirties",
+      "cer": 0.1499870923328457,
+      "wer": 0.38097560975609757
+    },
+    {
+      "subset_type": "age",
+      "subset": "fifties",
+      "cer": 0.19986216402481047,
+      "wer": 0.41885325558794945
+    },
+    {
+      "subset_type": "gender",
+      "subset": "",
+      "cer": 0.15221117167876436,
+      "wer": 0.3694843855740497
+    },
+    {
+      "subset_type": "gender",
+      "subset": "other",
+      "cer": 0.19473319473319473,
+      "wer": 0.42412451361867703
+    },
+    {
+      "subset_type": "gender",
+      "subset": "male",
+      "cer": 0.1668030483777149,
+      "wer": 0.404450188233972
+    },
+    {
+      "subset_type": "gender",
+      "subset": "female",
+      "cer": 0.15489725670178367,
+      "wer": 0.38953315492120133
+    },
+    {
+      "subset_type": "accent",
+      "subset": "",
+      "cer": 0.1545156412522236,
+      "wer": 0.37461315906899084
+    },
+    {
+      "subset_type": "accent",
+      "subset": "australia",
+      "cer": 0.1247323340471092,
+      "wer": 0.3630573248407643
+    },
+    {
+      "subset_type": "accent",
+      "subset": "singapore",
+      "cer": 0.19786096256684493,
+      "wer": 0.48484848484848486
+    },
+    {
+      "subset_type": "accent",
+      "subset": "wales",
+      "cer": 0.18493150684931506,
+      "wer": 0.4166666666666667
+    },
+    {
+      "subset_type": "accent",
+      "subset": "indian",
+      "cer": 0.18692307692307691,
+      "wer": 0.44172164624568017
+    },
+    {
+      "subset_type": "accent",
+      "subset": "scotland",
+      "cer": 0.11220196353436185,
+      "wer": 0.2907801418439716
+    },
+    {
+      "subset_type": "accent",
+      "subset": "england",
+      "cer": 0.13736535230754027,
+      "wer": 0.3617863199547767
+    },
+    {
+      "subset_type": "accent",
+      "subset": "ireland",
+      "cer": 0.1320754716981132,
+      "wer": 0.3206751054852321
+    },
+    {
+      "subset_type": "accent",
+      "subset": "us",
+      "cer": 0.13268065772392904,
+      "wer": 0.33684854042488155
+    },
+    {
+      "subset_type": "accent",
+      "subset": "newzealand",
+      "cer": 0.07681940700808626,
+      "wer": 0.27205882352941174
+    },
+    {
+      "subset_type": "accent",
+      "subset": "other",
+      "cer": 0.18642745709828393,
+      "wer": 0.3938223938223938
+    },
+    {
+      "subset_type": "accent",
+      "subset": "african",
+      "cer": 0.18655967903711135,
+      "wer": 0.4161849710982659
+    },
+    {
+      "subset_type": "accent",
+      "subset": "malaysia",
+      "cer": 0.11534154535274356,
+      "wer": 0.3248407643312102
+    },
+    {
+      "subset_type": "accent",
+      "subset": "hongkong",
+      "cer": 0.2846820809248555,
+      "wer": 0.65
+    },
+    {
+      "subset_type": "accent",
+      "subset": "bermuda",
+      "cer": 0.17303370786516853,
+      "wer": 0.4166666666666667
+    },
+    {
+      "subset_type": "accent",
+      "subset": "canada",
+      "cer": 0.16906653426017876,
+      "wer": 0.34269662921348315
+    },
+    {
+      "subset_type": "accent",
+      "subset": "philippines",
+      "cer": 0.10702341137123746,
+      "wer": 0.32323232323232326
+    }
+  ]
+}

cv7_transcriptions.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2901267e8e19ade993e557a3e6e0c7e21ac22f93288b6eff33ed72894a44417d
+size 122543347

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 sha256:8a6015b692764b3e8a8cbb74dee72458a4eaadcf9e67c945f5ed8768787b8c43
+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/tmp138cfn7s", "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}