jonatasgrosman/exp_w2v2r_en_xls-r_age_teens-5_sixties-5_s279

.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_s279
+
+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": 0.10696353874612474,
+      "wer": 0.30186296574143534
+    },
+    {
+      "subset_type": "age",
+      "subset": "",
+      "cer": 0.10549035253248427,
+      "wer": 0.2983692204351233
+    },
+    {
+      "subset_type": "age",
+      "subset": "sixties",
+      "cer": 0.07852154658490887,
+      "wer": 0.23376623376623376
+    },
+    {
+      "subset_type": "age",
+      "subset": "nineties",
+      "cer": 0.07692307692307693,
+      "wer": 0.2857142857142857
+    },
+    {
+      "subset_type": "age",
+      "subset": "fourties",
+      "cer": 0.09905422446406054,
+      "wer": 0.28717948717948716
+    },
+    {
+      "subset_type": "age",
+      "subset": "twenties",
+      "cer": 0.1281716006758246,
+      "wer": 0.3495571416286869
+    },
+    {
+      "subset_type": "age",
+      "subset": "teens",
+      "cer": 0.11143784222267289,
+      "wer": 0.31275720164609055
+    },
+    {
+      "subset_type": "age",
+      "subset": "eighties",
+      "cer": 0.0931899641577061,
+      "wer": 0.28846153846153844
+    },
+    {
+      "subset_type": "age",
+      "subset": "seventies",
+      "cer": 0.07209106239460371,
+      "wer": 0.25598086124401914
+    },
+    {
+      "subset_type": "age",
+      "subset": "thirties",
+      "cer": 0.11225838947302147,
+      "wer": 0.312922683238896
+    },
+    {
+      "subset_type": "age",
+      "subset": "fifties",
+      "cer": 0.08378305450620616,
+      "wer": 0.25770750988142294
+    },
+    {
+      "subset_type": "gender",
+      "subset": "",
+      "cer": 0.1055699515098648,
+      "wer": 0.2985842121368172
+    },
+    {
+      "subset_type": "gender",
+      "subset": "other",
+      "cer": 0.11572700296735905,
+      "wer": 0.32934131736526945
+    },
+    {
+      "subset_type": "gender",
+      "subset": "male",
+      "cer": 0.1151126751468974,
+      "wer": 0.32151422525276274
+    },
+    {
+      "subset_type": "gender",
+      "subset": "female",
+      "cer": 0.10928167243005345,
+      "wer": 0.30466605672461117
+    },
+    {
+      "subset_type": "accent",
+      "subset": "",
+      "cer": 0.10778820023637077,
+      "wer": 0.3032250386766631
+    },
+    {
+      "subset_type": "accent",
+      "subset": "australia",
+      "cer": 0.06726794438181134,
+      "wer": 0.22171945701357465
+    },
+    {
+      "subset_type": "accent",
+      "subset": "wales",
+      "cer": 0.08248587570621468,
+      "wer": 0.2638888888888889
+    },
+    {
+      "subset_type": "accent",
+      "subset": "singapore",
+      "cer": 0.16132723112128147,
+      "wer": 0.3116883116883117
+    },
+    {
+      "subset_type": "accent",
+      "subset": "england",
+      "cer": 0.09859154929577464,
+      "wer": 0.2929493545183714
+    },
+    {
+      "subset_type": "accent",
+      "subset": "indian",
+      "cer": 0.13812268423219431,
+      "wer": 0.3710162357185809
+    },
+    {
+      "subset_type": "accent",
+      "subset": "scotland",
+      "cer": 0.07295373665480427,
+      "wer": 0.2840909090909091
+    },
+    {
+      "subset_type": "accent",
+      "subset": "southatlandtic",
+      "cer": 0.14285714285714285,
+      "wer": 0.36363636363636365
+    },
+    {
+      "subset_type": "accent",
+      "subset": "ireland",
+      "cer": 0.11264822134387352,
+      "wer": 0.31496062992125984
+    },
+    {
+      "subset_type": "accent",
+      "subset": "us",
+      "cer": 0.08645260823450004,
+      "wer": 0.2605999097880018
+    },
+    {
+      "subset_type": "accent",
+      "subset": "newzealand",
+      "cer": 0.04926108374384237,
+      "wer": 0.2
+    },
+    {
+      "subset_type": "accent",
+      "subset": "other",
+      "cer": 0.10408921933085502,
+      "wer": 0.32
+    },
+    {
+      "subset_type": "accent",
+      "subset": "african",
+      "cer": 0.12884333821376281,
+      "wer": 0.3673469387755102
+    },
+    {
+      "subset_type": "accent",
+      "subset": "malaysia",
+      "cer": 0.15128205128205127,
+      "wer": 0.3333333333333333
+    },
+    {
+      "subset_type": "accent",
+      "subset": "hongkong",
+      "cer": 0.09392265193370165,
+      "wer": 0.2826086956521739
+    },
+    {
+      "subset_type": "accent",
+      "subset": "bermuda",
+      "cer": 0.06172839506172839,
+      "wer": 0.125
+    },
+    {
+      "subset_type": "accent",
+      "subset": "canada",
+      "cer": 0.07230142566191446,
+      "wer": 0.23621103117505995
+    },
+    {
+      "subset_type": "accent",
+      "subset": "philippines",
+      "cer": 0.15823914336704342,
+      "wer": 0.4117647058823529
+    }
+  ],
+  "test": [
+    {
+      "subset_type": "all",
+      "subset": "all",
+      "cer": 0.13912906812268253,
+      "wer": 0.3496632446217224
+    },
+    {
+      "subset_type": "age",
+      "subset": "",
+      "cer": 0.13770383635255287,
+      "wer": 0.3463764918873278
+    },
+    {
+      "subset_type": "age",
+      "subset": "sixties",
+      "cer": 0.14113873295910184,
+      "wer": 0.28854625550660795
+    },
+    {
+      "subset_type": "age",
+      "subset": "fourties",
+      "cer": 0.132195080720142,
+      "wer": 0.32847744360902253
+    },
+    {
+      "subset_type": "age",
+      "subset": "twenties",
+      "cer": 0.15253227408143,
+      "wer": 0.3821091740497285
+    },
+    {
+      "subset_type": "age",
+      "subset": "teens",
+      "cer": 0.14683272748942255,
+      "wer": 0.37678692988427503
+    },
+    {
+      "subset_type": "age",
+      "subset": "eighties",
+      "cer": 0.11428571428571428,
+      "wer": 0.48
+    },
+    {
+      "subset_type": "age",
+      "subset": "seventies",
+      "cer": 0.10427968655816756,
+      "wer": 0.27181208053691275
+    },
+    {
+      "subset_type": "age",
+      "subset": "thirties",
+      "cer": 0.1369073229498322,
+      "wer": 0.3595121951219512
+    },
+    {
+      "subset_type": "age",
+      "subset": "fifties",
+      "cer": 0.1910751206064783,
+      "wer": 0.41205053449951407
+    },
+    {
+      "subset_type": "gender",
+      "subset": "",
+      "cer": 0.13765322281885206,
+      "wer": 0.3463155733653707
+    },
+    {
+      "subset_type": "gender",
+      "subset": "other",
+      "cer": 0.1864171864171864,
+      "wer": 0.4046692607003891
+    },
+    {
+      "subset_type": "gender",
+      "subset": "male",
+      "cer": 0.14946121008370222,
+      "wer": 0.37393942799348207
+    },
+    {
+      "subset_type": "gender",
+      "subset": "female",
+      "cer": 0.13763429644310002,
+      "wer": 0.3479036574487065
+    },
+    {
+      "subset_type": "accent",
+      "subset": "",
+      "cer": 0.13953169227691128,
+      "wer": 0.35036632936536494
+    },
+    {
+      "subset_type": "accent",
+      "subset": "australia",
+      "cer": 0.0968950749464668,
+      "wer": 0.28662420382165604
+    },
+    {
+      "subset_type": "accent",
+      "subset": "singapore",
+      "cer": 0.13368983957219252,
+      "wer": 0.3939393939393939
+    },
+    {
+      "subset_type": "accent",
+      "subset": "wales",
+      "cer": 0.18493150684931506,
+      "wer": 0.375
+    },
+    {
+      "subset_type": "accent",
+      "subset": "indian",
+      "cer": 0.15697802197802196,
+      "wer": 0.38894125039271127
+    },
+    {
+      "subset_type": "accent",
+      "subset": "scotland",
+      "cer": 0.1388499298737728,
+      "wer": 0.3333333333333333
+    },
+    {
+      "subset_type": "accent",
+      "subset": "england",
+      "cer": 0.12343116908785454,
+      "wer": 0.33578292820802713
+    },
+    {
+      "subset_type": "accent",
+      "subset": "ireland",
+      "cer": 0.1081277213352685,
+      "wer": 0.3080168776371308
+    },
+    {
+      "subset_type": "accent",
+      "subset": "us",
+      "cer": 0.12297165729121592,
+      "wer": 0.32156503133119363
+    },
+    {
+      "subset_type": "accent",
+      "subset": "newzealand",
+      "cer": 0.10377358490566038,
+      "wer": 0.3161764705882353
+    },
+    {
+      "subset_type": "accent",
+      "subset": "other",
+      "cer": 0.2059282371294852,
+      "wer": 0.4247104247104247
+    },
+    {
+      "subset_type": "accent",
+      "subset": "african",
+      "cer": 0.17552657973921765,
+      "wer": 0.3930635838150289
+    },
+    {
+      "subset_type": "accent",
+      "subset": "malaysia",
+      "cer": 0.11310190369540873,
+      "wer": 0.2929936305732484
+    },
+    {
+      "subset_type": "accent",
+      "subset": "hongkong",
+      "cer": 0.24132947976878613,
+      "wer": 0.55
+    },
+    {
+      "subset_type": "accent",
+      "subset": "bermuda",
+      "cer": 0.11910112359550562,
+      "wer": 0.2857142857142857
+    },
+    {
+      "subset_type": "accent",
+      "subset": "canada",
+      "cer": 0.16782522343594836,
+      "wer": 0.3300561797752809
+    },
+    {
+      "subset_type": "accent",
+      "subset": "philippines",
+      "cer": 0.11204013377926421,
+      "wer": 0.32323232323232326
+    }
+  ]
+}

cv7_transcriptions.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:faec70009d6e332155a19c1a1c8158e3399fb9ef3ac85b1a981cce58223bd911
+size 123270535

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 sha256:5a908e5e5dc94aa443f38aa8fa1f1f5ecca1c87caa5f88be45225322cfec3942
+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/tmpmt5ksbsj", "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}