jonatasgrosman/exp_w2v2r_en_xls-r_age_teens-10_sixties-0_s807

.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-10_sixties-0_s807
+
+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": 0.10200063164272224,
+      "wer": 0.28917854463615905
+    },
+    {
+      "subset_type": "age",
+      "subset": "",
+      "cer": 0.10081446408160805,
+      "wer": 0.2866365637944546
+    },
+    {
+      "subset_type": "age",
+      "subset": "sixties",
+      "cer": 0.08056549139839891,
+      "wer": 0.23976023976023977
+    },
+    {
+      "subset_type": "age",
+      "subset": "nineties",
+      "cer": 0.038461538461538464,
+      "wer": 0.14285714285714285
+    },
+    {
+      "subset_type": "age",
+      "subset": "fourties",
+      "cer": 0.09363177805800757,
+      "wer": 0.2714285714285714
+    },
+    {
+      "subset_type": "age",
+      "subset": "twenties",
+      "cer": 0.12022331594799089,
+      "wer": 0.32874709777280936
+    },
+    {
+      "subset_type": "age",
+      "subset": "teens",
+      "cer": 0.10271750152098966,
+      "wer": 0.2854203409758965
+    },
+    {
+      "subset_type": "age",
+      "subset": "eighties",
+      "cer": 0.11469534050179211,
+      "wer": 0.3269230769230769
+    },
+    {
+      "subset_type": "age",
+      "subset": "seventies",
+      "cer": 0.081787521079258,
+      "wer": 0.24641148325358853
+    },
+    {
+      "subset_type": "age",
+      "subset": "thirties",
+      "cer": 0.10681231505383114,
+      "wer": 0.301041857064522
+    },
+    {
+      "subset_type": "age",
+      "subset": "fifties",
+      "cer": 0.0779816513761468,
+      "wer": 0.24426877470355732
+    },
+    {
+      "subset_type": "gender",
+      "subset": "",
+      "cer": 0.10087744054185802,
+      "wer": 0.28672502966661445
+    },
+    {
+      "subset_type": "gender",
+      "subset": "other",
+      "cer": 0.09099901088031652,
+      "wer": 0.2754491017964072
+    },
+    {
+      "subset_type": "gender",
+      "subset": "male",
+      "cer": 0.10841350810357074,
+      "wer": 0.3029861274394545
+    },
+    {
+      "subset_type": "gender",
+      "subset": "female",
+      "cer": 0.1060594152782144,
+      "wer": 0.2982616651418115
+    },
+    {
+      "subset_type": "accent",
+      "subset": "",
+      "cer": 0.10281286030294934,
+      "wer": 0.29063149715437764
+    },
+    {
+      "subset_type": "accent",
+      "subset": "australia",
+      "cer": 0.05787298008267568,
+      "wer": 0.20135746606334842
+    },
+    {
+      "subset_type": "accent",
+      "subset": "wales",
+      "cer": 0.07909604519774012,
+      "wer": 0.2847222222222222
+    },
+    {
+      "subset_type": "accent",
+      "subset": "singapore",
+      "cer": 0.16247139588100687,
+      "wer": 0.3051948051948052
+    },
+    {
+      "subset_type": "accent",
+      "subset": "england",
+      "cer": 0.09197870147715562,
+      "wer": 0.2845084409136048
+    },
+    {
+      "subset_type": "accent",
+      "subset": "indian",
+      "cer": 0.1288596130094689,
+      "wer": 0.33944678292242936
+    },
+    {
+      "subset_type": "accent",
+      "subset": "scotland",
+      "cer": 0.0800711743772242,
+      "wer": 0.23863636363636365
+    },
+    {
+      "subset_type": "accent",
+      "subset": "southatlandtic",
+      "cer": 0.13736263736263737,
+      "wer": 0.3939393939393939
+    },
+    {
+      "subset_type": "accent",
+      "subset": "ireland",
+      "cer": 0.10013175230566534,
+      "wer": 0.2874015748031496
+    },
+    {
+      "subset_type": "accent",
+      "subset": "us",
+      "cer": 0.08330922765403528,
+      "wer": 0.2523680649526387
+    },
+    {
+      "subset_type": "accent",
+      "subset": "newzealand",
+      "cer": 0.04433497536945813,
+      "wer": 0.2
+    },
+    {
+      "subset_type": "accent",
+      "subset": "other",
+      "cer": 0.08736059479553904,
+      "wer": 0.34
+    },
+    {
+      "subset_type": "accent",
+      "subset": "african",
+      "cer": 0.12347486578818936,
+      "wer": 0.33527696793002915
+    },
+    {
+      "subset_type": "accent",
+      "subset": "malaysia",
+      "cer": 0.13846153846153847,
+      "wer": 0.3472222222222222
+    },
+    {
+      "subset_type": "accent",
+      "subset": "hongkong",
+      "cer": 0.09760589318600368,
+      "wer": 0.2608695652173913
+    },
+    {
+      "subset_type": "accent",
+      "subset": "bermuda",
+      "cer": 0.06172839506172839,
+      "wer": 0.125
+    },
+    {
+      "subset_type": "accent",
+      "subset": "canada",
+      "cer": 0.07230142566191446,
+      "wer": 0.22661870503597123
+    },
+    {
+      "subset_type": "accent",
+      "subset": "philippines",
+      "cer": 0.1588340273646639,
+      "wer": 0.4117647058823529
+    }
+  ],
+  "test": [
+    {
+      "subset_type": "all",
+      "subset": "all",
+      "cer": 0.1319995576071202,
+      "wer": 0.3330020270712091
+    },
+    {
+      "subset_type": "age",
+      "subset": "",
+      "cer": 0.13089786425798883,
+      "wer": 0.32983926791976326
+    },
+    {
+      "subset_type": "age",
+      "subset": "sixties",
+      "cer": 0.1351242983159583,
+      "wer": 0.2907488986784141
+    },
+    {
+      "subset_type": "age",
+      "subset": "fourties",
+      "cer": 0.12779984785732398,
+      "wer": 0.32095864661654133
+    },
+    {
+      "subset_type": "age",
+      "subset": "twenties",
+      "cer": 0.1410956636875207,
+      "wer": 0.362770315328189
+    },
+    {
+      "subset_type": "age",
+      "subset": "teens",
+      "cer": 0.1446278529289077,
+      "wer": 0.3713410483321988
+    },
+    {
+      "subset_type": "age",
+      "subset": "eighties",
+      "cer": 0.12,
+      "wer": 0.44
+    },
+    {
+      "subset_type": "age",
+      "subset": "seventies",
+      "cer": 0.10427968655816756,
+      "wer": 0.2651006711409396
+    },
+    {
+      "subset_type": "age",
+      "subset": "thirties",
+      "cer": 0.12584975475432406,
+      "wer": 0.3353658536585366
+    },
+    {
+      "subset_type": "age",
+      "subset": "fifties",
+      "cer": 0.18073742246726396,
+      "wer": 0.3751214771622935
+    },
+    {
+      "subset_type": "gender",
+      "subset": "",
+      "cer": 0.13086955518735488,
+      "wer": 0.32982290724110924
+    },
+    {
+      "subset_type": "gender",
+      "subset": "other",
+      "cer": 0.17325017325017325,
+      "wer": 0.4163424124513619
+    },
+    {
+      "subset_type": "gender",
+      "subset": "male",
+      "cer": 0.13956285553726178,
+      "wer": 0.3543293813564084
+    },
+    {
+      "subset_type": "gender",
+      "subset": "female",
+      "cer": 0.13231459267758422,
+      "wer": 0.33809099018733274
+    },
+    {
+      "subset_type": "accent",
+      "subset": "",
+      "cer": 0.13265084949230538,
+      "wer": 0.3339642738905762
+    },
+    {
+      "subset_type": "accent",
+      "subset": "australia",
+      "cer": 0.10278372591006424,
+      "wer": 0.2929936305732484
+    },
+    {
+      "subset_type": "accent",
+      "subset": "singapore",
+      "cer": 0.12299465240641712,
+      "wer": 0.3333333333333333
+    },
+    {
+      "subset_type": "accent",
+      "subset": "wales",
+      "cer": 0.1780821917808219,
+      "wer": 0.3333333333333333
+    },
+    {
+      "subset_type": "accent",
+      "subset": "indian",
+      "cer": 0.14445054945054944,
+      "wer": 0.3644360666038329
+    },
+    {
+      "subset_type": "accent",
+      "subset": "scotland",
+      "cer": 0.12061711079943899,
+      "wer": 0.2907801418439716
+    },
+    {
+      "subset_type": "accent",
+      "subset": "england",
+      "cer": 0.11404288961359818,
+      "wer": 0.3109101187111362
+    },
+    {
+      "subset_type": "accent",
+      "subset": "ireland",
+      "cer": 0.10087082728592163,
+      "wer": 0.29957805907172996
+    },
+    {
+      "subset_type": "accent",
+      "subset": "us",
+      "cer": 0.11450670705322372,
+      "wer": 0.3065871924193795
+    },
+    {
+      "subset_type": "accent",
+      "subset": "newzealand",
+      "cer": 0.09164420485175202,
+      "wer": 0.3014705882352941
+    },
+    {
+      "subset_type": "accent",
+      "subset": "other",
+      "cer": 0.1856474258970359,
+      "wer": 0.38223938223938225
+    },
+    {
+      "subset_type": "accent",
+      "subset": "african",
+      "cer": 0.15947843530591777,
+      "wer": 0.3872832369942196
+    },
+    {
+      "subset_type": "accent",
+      "subset": "malaysia",
+      "cer": 0.09742441209406495,
+      "wer": 0.3184713375796178
+    },
+    {
+      "subset_type": "accent",
+      "subset": "hongkong",
+      "cer": 0.22976878612716764,
+      "wer": 0.575
+    },
+    {
+      "subset_type": "accent",
+      "subset": "bermuda",
+      "cer": 0.12134831460674157,
+      "wer": 0.27380952380952384
+    },
+    {
+      "subset_type": "accent",
+      "subset": "canada",
+      "cer": 0.15938430983118174,
+      "wer": 0.29353932584269665
+    },
+    {
+      "subset_type": "accent",
+      "subset": "philippines",
+      "cer": 0.10869565217391304,
+      "wer": 0.26262626262626265
+    }
+  ]
+}

cv7_transcriptions.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a6adcc314a8a75c25fce952f73e4db675f9de3aa5547a7790f460078fb4742b6
+size 123405798

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 sha256:9f2762eba9aebf08802fc8e5b49113a4bb21eb9ca1bdae3ab77622467d0a86a9
+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/tmptk2sbyg4", "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}