jonatasgrosman/exp_w2v2r_en_xls-r_gender_male-0_female-10_s727

.gitattributes

@@ -29,3 +29,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zstandard 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_gender_male-0_female-10_s727
+
+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": 0.10672720983643891,
+      "wer": 0.30180045011252815
+    },
+    {
+      "subset_type": "age",
+      "subset": "",
+      "cer": 0.1053299997049508,
+      "wer": 0.29843639213344775
+    },
+    {
+      "subset_type": "age",
+      "subset": "sixties",
+      "cer": 0.08380173735309146,
+      "wer": 0.24675324675324675
+    },
+    {
+      "subset_type": "age",
+      "subset": "seventies",
+      "cer": 0.08136593591905565,
+      "wer": 0.24880382775119617
+    },
+    {
+      "subset_type": "age",
+      "subset": "fourties",
+      "cer": 0.09886506935687264,
+      "wer": 0.2904761904761905
+    },
+    {
+      "subset_type": "age",
+      "subset": "thirties",
+      "cer": 0.11269911225838947,
+      "wer": 0.315481630414915
+    },
+    {
+      "subset_type": "age",
+      "subset": "fifties",
+      "cer": 0.08000539665407447,
+      "wer": 0.25770750988142294
+    },
+    {
+      "subset_type": "age",
+      "subset": "teens",
+      "cer": 0.11387142567430542,
+      "wer": 0.32686654908877133
+    },
+    {
+      "subset_type": "age",
+      "subset": "twenties",
+      "cer": 0.1254683023580401,
+      "wer": 0.34095794995270445
+    },
+    {
+      "subset_type": "age",
+      "subset": "nineties",
+      "cer": 0.0,
+      "wer": 0.0
+    },
+    {
+      "subset_type": "age",
+      "subset": "eighties",
+      "cer": 0.11827956989247312,
+      "wer": 0.3269230769230769
+    },
+    {
+      "subset_type": "gender",
+      "subset": "",
+      "cer": 0.10538394437745337,
+      "wer": 0.29852450574300876
+    },
+    {
+      "subset_type": "gender",
+      "subset": "female",
+      "cer": 0.10613800691606413,
+      "wer": 0.30146386093321137
+    },
+    {
+      "subset_type": "gender",
+      "subset": "male",
+      "cer": 0.11534674242913412,
+      "wer": 0.32240771220315073
+    },
+    {
+      "subset_type": "gender",
+      "subset": "other",
+      "cer": 0.10385756676557864,
+      "wer": 0.3083832335329341
+    },
+    {
+      "subset_type": "accent",
+      "subset": "",
+      "cer": 0.10749093149821037,
+      "wer": 0.30300802934526183
+    },
+    {
+      "subset_type": "accent",
+      "subset": "african",
+      "cer": 0.1273792093704246,
+      "wer": 0.358600583090379
+    },
+    {
+      "subset_type": "accent",
+      "subset": "bermuda",
+      "cer": 0.037037037037037035,
+      "wer": 0.125
+    },
+    {
+      "subset_type": "accent",
+      "subset": "scotland",
+      "cer": 0.07829181494661921,
+      "wer": 0.2159090909090909
+    },
+    {
+      "subset_type": "accent",
+      "subset": "malaysia",
+      "cer": 0.14358974358974358,
+      "wer": 0.3888888888888889
+    },
+    {
+      "subset_type": "accent",
+      "subset": "wales",
+      "cer": 0.06892655367231638,
+      "wer": 0.2638888888888889
+    },
+    {
+      "subset_type": "accent",
+      "subset": "canada",
+      "cer": 0.07270875763747454,
+      "wer": 0.2290167865707434
+    },
+    {
+      "subset_type": "accent",
+      "subset": "us",
+      "cer": 0.08822678623083599,
+      "wer": 0.2668019846639603
+    },
+    {
+      "subset_type": "accent",
+      "subset": "hongkong",
+      "cer": 0.10313075506445672,
+      "wer": 0.30434782608695654
+    },
+    {
+      "subset_type": "accent",
+      "subset": "newzealand",
+      "cer": 0.03940886699507389,
+      "wer": 0.15
+    },
+    {
+      "subset_type": "accent",
+      "subset": "ireland",
+      "cer": 0.11067193675889328,
+      "wer": 0.3425196850393701
+    },
+    {
+      "subset_type": "accent",
+      "subset": "southatlandtic",
+      "cer": 0.15384615384615385,
+      "wer": 0.3939393939393939
+    },
+    {
+      "subset_type": "accent",
+      "subset": "other",
+      "cer": 0.08178438661710037,
+      "wer": 0.29
+    },
+    {
+      "subset_type": "accent",
+      "subset": "indian",
+      "cer": 0.1380712227254014,
+      "wer": 0.3619963920625376
+    },
+    {
+      "subset_type": "accent",
+      "subset": "england",
+      "cer": 0.0956715905187221,
+      "wer": 0.2869910625620655
+    },
+    {
+      "subset_type": "accent",
+      "subset": "australia",
+      "cer": 0.06952273581360391,
+      "wer": 0.2398190045248869
+    },
+    {
+      "subset_type": "accent",
+      "subset": "singapore",
+      "cer": 0.16475972540045766,
+      "wer": 0.33116883116883117
+    },
+    {
+      "subset_type": "accent",
+      "subset": "philippines",
+      "cer": 0.14991076740035694,
+      "wer": 0.42214532871972316
+    }
+  ],
+  "test": [
+    {
+      "subset_type": "all",
+      "subset": "all",
+      "cer": 0.13802541430675289,
+      "wer": 0.3478388805335775
+    },
+    {
+      "subset_type": "age",
+      "subset": "",
+      "cer": 0.1368838069998466,
+      "wer": 0.3446345303930443
+    },
+    {
+      "subset_type": "age",
+      "subset": "seventies",
+      "cer": 0.10849909584086799,
+      "wer": 0.28859060402684567
+    },
+    {
+      "subset_type": "age",
+      "subset": "sixties",
+      "cer": 0.15036086607858862,
+      "wer": 0.31718061674008813
+    },
+    {
+      "subset_type": "age",
+      "subset": "fourties",
+      "cer": 0.12923675090862988,
+      "wer": 0.33270676691729323
+    },
+    {
+      "subset_type": "age",
+      "subset": "thirties",
+      "cer": 0.13200240943120214,
+      "wer": 0.35536585365853657
+    },
+    {
+      "subset_type": "age",
+      "subset": "fifties",
+      "cer": 0.19297036526533426,
+      "wer": 0.3945578231292517
+    },
+    {
+      "subset_type": "age",
+      "subset": "teens",
+      "cer": 0.14963351409331982,
+      "wer": 0.37440435670524164
+    },
+    {
+      "subset_type": "age",
+      "subset": "twenties",
+      "cer": 0.14774908970539558,
+      "wer": 0.37867962274935696
+    },
+    {
+      "subset_type": "age",
+      "subset": "eighties",
+      "cer": 0.14857142857142858,
+      "wer": 0.52
+    },
+    {
+      "subset_type": "gender",
+      "subset": "",
+      "cer": 0.13681730778364457,
+      "wer": 0.3445514892063902
+    },
+    {
+      "subset_type": "gender",
+      "subset": "female",
+      "cer": 0.13768645040158548,
+      "wer": 0.35622955694320546
+    },
+    {
+      "subset_type": "gender",
+      "subset": "male",
+      "cer": 0.14616833118080272,
+      "wer": 0.36916334213631513
+    },
+    {
+      "subset_type": "gender",
+      "subset": "other",
+      "cer": 0.18711018711018712,
+      "wer": 0.44357976653696496
+    },
+    {
+      "subset_type": "accent",
+      "subset": "",
+      "cer": 0.13846333402163136,
+      "wer": 0.3482360053545982
+    },
+    {
+      "subset_type": "accent",
+      "subset": "african",
+      "cer": 0.17753259779338015,
+      "wer": 0.4046242774566474
+    },
+    {
+      "subset_type": "accent",
+      "subset": "bermuda",
+      "cer": 0.14606741573033707,
+      "wer": 0.34523809523809523
+    },
+    {
+      "subset_type": "accent",
+      "subset": "scotland",
+      "cer": 0.10659186535764376,
+      "wer": 0.2765957446808511
+    },
+    {
+      "subset_type": "accent",
+      "subset": "malaysia",
+      "cer": 0.11310190369540873,
+      "wer": 0.3248407643312102
+    },
+    {
+      "subset_type": "accent",
+      "subset": "wales",
+      "cer": 0.19863013698630136,
+      "wer": 0.375
+    },
+    {
+      "subset_type": "accent",
+      "subset": "canada",
+      "cer": 0.1663356504468719,
+      "wer": 0.33707865168539325
+    },
+    {
+      "subset_type": "accent",
+      "subset": "us",
+      "cer": 0.12226849848550411,
+      "wer": 0.3238575576952468
+    },
+    {
+      "subset_type": "accent",
+      "subset": "hongkong",
+      "cer": 0.2514450867052023,
+      "wer": 0.5583333333333333
+    },
+    {
+      "subset_type": "accent",
+      "subset": "newzealand",
+      "cer": 0.07142857142857142,
+      "wer": 0.29411764705882354
+    },
+    {
+      "subset_type": "accent",
+      "subset": "ireland",
+      "cer": 0.11973875181422351,
+      "wer": 0.31645569620253167
+    },
+    {
+      "subset_type": "accent",
+      "subset": "other",
+      "cer": 0.19656786271450857,
+      "wer": 0.40540540540540543
+    },
+    {
+      "subset_type": "accent",
+      "subset": "indian",
+      "cer": 0.1478021978021978,
+      "wer": 0.3776311655670751
+    },
+    {
+      "subset_type": "accent",
+      "subset": "england",
+      "cer": 0.12837236881114736,
+      "wer": 0.34765404183154325
+    },
+    {
+      "subset_type": "accent",
+      "subset": "australia",
+      "cer": 0.11723768736616702,
+      "wer": 0.3248407643312102
+    },
+    {
+      "subset_type": "accent",
+      "subset": "singapore",
+      "cer": 0.16042780748663102,
+      "wer": 0.42424242424242425
+    },
+    {
+      "subset_type": "accent",
+      "subset": "philippines",
+      "cer": 0.13043478260869565,
+      "wer": 0.32323232323232326
+    }
+  ]
+}

cv7_transcriptions.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:84ee2e52ed832b89e57a0a8fdba1909ef6d4c794c145984e9310ea30c909e8ba
+size 123164005

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 sha256:0d1c61604ae5fca2f3c6f94420e2e6b4d00f566cd9879d785985ee07579e9623
+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/tmp_bbosuka", "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}