jonatasgrosman/exp_w2v2r_en_vp-100k_age_teens-5_sixties-5_s649

.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_vp-100k_age_teens-5_sixties-5_s649
+
+Fine-tuned [facebook/wav2vec2-large-100k-voxpopuli](https://huggingface.co/facebook/wav2vec2-large-100k-voxpopuli) 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 +1,116 @@
+{
+  "_name_or_path": "facebook/wav2vec2-large-100k-voxpopuli",
+  "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_channel_length": 10,
+  "mask_channel_min_space": 1,
+  "mask_channel_other": 0.0,
+  "mask_channel_prob": 0.0,
+  "mask_channel_selection": "static",
+  "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_min_space": 1,
+  "mask_time_other": 0.0,
+  "mask_time_prob": 0.05,
+  "mask_time_selection": "static",
+  "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": 0.15369972908113172,
+      "wer": 0.4150350087521881
+    },
+    {
+      "subset_type": "age",
+      "subset": "",
+      "cer": 0.15225565115434794,
+      "wer": 0.41193417173564206
+    },
+    {
+      "subset_type": "age",
+      "subset": "thirties",
+      "cer": 0.15777875716174528,
+      "wer": 0.42277462986656916
+    },
+    {
+      "subset_type": "age",
+      "subset": "sixties",
+      "cer": 0.11650485436893204,
+      "wer": 0.34865134865134867
+    },
+    {
+      "subset_type": "age",
+      "subset": "seventies",
+      "cer": 0.13659359190556492,
+      "wer": 0.3660287081339713
+    },
+    {
+      "subset_type": "age",
+      "subset": "fifties",
+      "cer": 0.13100377765785212,
+      "wer": 0.3778656126482213
+    },
+    {
+      "subset_type": "age",
+      "subset": "twenties",
+      "cer": 0.17639021523543671,
+      "wer": 0.458336916329865
+    },
+    {
+      "subset_type": "age",
+      "subset": "eighties",
+      "cer": 0.17562724014336917,
+      "wer": 0.4230769230769231
+    },
+    {
+      "subset_type": "age",
+      "subset": "nineties",
+      "cer": 0.23076923076923078,
+      "wer": 0.5714285714285714
+    },
+    {
+      "subset_type": "age",
+      "subset": "teens",
+      "cer": 0.15828432366659906,
+      "wer": 0.43356848912404466
+    },
+    {
+      "subset_type": "age",
+      "subset": "fourties",
+      "cer": 0.13984867591424968,
+      "wer": 0.39267399267399267
+    },
+    {
+      "subset_type": "gender",
+      "subset": "",
+      "cer": 0.15234112425276444,
+      "wer": 0.4121233832628052
+    },
+    {
+      "subset_type": "gender",
+      "subset": "male",
+      "cer": 0.16125621488990766,
+      "wer": 0.43080178697390076
+    },
+    {
+      "subset_type": "gender",
+      "subset": "other",
+      "cer": 0.14985163204747776,
+      "wer": 0.4281437125748503
+    },
+    {
+      "subset_type": "gender",
+      "subset": "female",
+      "cer": 0.15883370009430997,
+      "wer": 0.4265782250686185
+    },
+    {
+      "subset_type": "accent",
+      "subset": "",
+      "cer": 0.15465556542439626,
+      "wer": 0.4161748955205074
+    },
+    {
+      "subset_type": "accent",
+      "subset": "england",
+      "cer": 0.14230504981106148,
+      "wer": 0.41062562065541214
+    },
+    {
+      "subset_type": "accent",
+      "subset": "australia",
+      "cer": 0.11086057872980083,
+      "wer": 0.33031674208144796
+    },
+    {
+      "subset_type": "accent",
+      "subset": "malaysia",
+      "cer": 0.19230769230769232,
+      "wer": 0.4166666666666667
+    },
+    {
+      "subset_type": "accent",
+      "subset": "indian",
+      "cer": 0.19076780568135035,
+      "wer": 0.4783523752254961
+    },
+    {
+      "subset_type": "accent",
+      "subset": "wales",
+      "cer": 0.09830508474576272,
+      "wer": 0.3611111111111111
+    },
+    {
+      "subset_type": "accent",
+      "subset": "philippines",
+      "cer": 0.206424747174301,
+      "wer": 0.5259515570934256
+    },
+    {
+      "subset_type": "accent",
+      "subset": "hongkong",
+      "cer": 0.16758747697974216,
+      "wer": 0.44565217391304346
+    },
+    {
+      "subset_type": "accent",
+      "subset": "bermuda",
+      "cer": 0.06172839506172839,
+      "wer": 0.125
+    },
+    {
+      "subset_type": "accent",
+      "subset": "ireland",
+      "cer": 0.1772068511198946,
+      "wer": 0.47244094488188976
+    },
+    {
+      "subset_type": "accent",
+      "subset": "newzealand",
+      "cer": 0.07881773399014778,
+      "wer": 0.275
+    },
+    {
+      "subset_type": "accent",
+      "subset": "other",
+      "cer": 0.1524163568773234,
+      "wer": 0.49
+    },
+    {
+      "subset_type": "accent",
+      "subset": "us",
+      "cer": 0.1296885546234693,
+      "wer": 0.37686062246278756
+    },
+    {
+      "subset_type": "accent",
+      "subset": "african",
+      "cer": 0.17130307467057102,
+      "wer": 0.44314868804664725
+    },
+    {
+      "subset_type": "accent",
+      "subset": "scotland",
+      "cer": 0.10498220640569395,
+      "wer": 0.375
+    },
+    {
+      "subset_type": "accent",
+      "subset": "singapore",
+      "cer": 0.2048054919908467,
+      "wer": 0.44155844155844154
+    },
+    {
+      "subset_type": "accent",
+      "subset": "southatlandtic",
+      "cer": 0.17582417582417584,
+      "wer": 0.5151515151515151
+    },
+    {
+      "subset_type": "accent",
+      "subset": "canada",
+      "cer": 0.12219959266802444,
+      "wer": 0.3657074340527578
+    }
+  ],
+  "test": [
+    {
+      "subset_type": "all",
+      "subset": "all",
+      "cer": 0.1950032888418038,
+      "wer": 0.47598901458183485
+    },
+    {
+      "subset_type": "age",
+      "subset": "",
+      "cer": 0.19396952586282973,
+      "wer": 0.4738211332638577
+    },
+    {
+      "subset_type": "age",
+      "subset": "thirties",
+      "cer": 0.18776353153773342,
+      "wer": 0.4714634146341463
+    },
+    {
+      "subset_type": "age",
+      "subset": "seventies",
+      "cer": 0.16033755274261605,
+      "wer": 0.40604026845637586
+    },
+    {
+      "subset_type": "age",
+      "subset": "fifties",
+      "cer": 0.25103376981392145,
+      "wer": 0.5519922254616132
+    },
+    {
+      "subset_type": "age",
+      "subset": "twenties",
+      "cer": 0.2006454816285998,
+      "wer": 0.4937601219396018
+    },
+    {
+      "subset_type": "age",
+      "subset": "eighties",
+      "cer": 0.13142857142857142,
+      "wer": 0.4
+    },
+    {
+      "subset_type": "age",
+      "subset": "fourties",
+      "cer": 0.19136167695038458,
+      "wer": 0.45864661654135336
+    },
+    {
+      "subset_type": "age",
+      "subset": "teens",
+      "cer": 0.21542220368273643,
+      "wer": 0.5136147038801906
+    },
+    {
+      "subset_type": "age",
+      "subset": "sixties",
+      "cer": 0.20809943865276664,
+      "wer": 0.44933920704845814
+    },
+    {
+      "subset_type": "gender",
+      "subset": "",
+      "cer": 0.19389700604002894,
+      "wer": 0.4737022195524397
+    },
+    {
+      "subset_type": "gender",
+      "subset": "male",
+      "cer": 0.20413877414202758,
+      "wer": 0.494184413103332
+    },
+    {
+      "subset_type": "gender",
+      "subset": "other",
+      "cer": 0.21968121968121967,
+      "wer": 0.5058365758754864
+    },
+    {
+      "subset_type": "gender",
+      "subset": "female",
+      "cer": 0.18733701887973297,
+      "wer": 0.46684507879869164
+    },
+    {
+      "subset_type": "accent",
+      "subset": "",
+      "cer": 0.19574426607122475,
+      "wer": 0.4771925784117571
+    },
+    {
+      "subset_type": "accent",
+      "subset": "england",
+      "cer": 0.1696807984978753,
+      "wer": 0.4431882419446015
+    },
+    {
+      "subset_type": "accent",
+      "subset": "malaysia",
+      "cer": 0.15789473684210525,
+      "wer": 0.4140127388535032
+    },
+    {
+      "subset_type": "accent",
+      "subset": "australia",
+      "cer": 0.16595289079229122,
+      "wer": 0.42038216560509556
+    },
+    {
+      "subset_type": "accent",
+      "subset": "indian",
+      "cer": 0.21137362637362636,
+      "wer": 0.5067546339930883
+    },
+    {
+      "subset_type": "accent",
+      "subset": "wales",
+      "cer": 0.17123287671232876,
+      "wer": 0.4166666666666667
+    },
+    {
+      "subset_type": "accent",
+      "subset": "philippines",
+      "cer": 0.15217391304347827,
+      "wer": 0.3838383838383838
+    },
+    {
+      "subset_type": "accent",
+      "subset": "hongkong",
+      "cer": 0.30346820809248554,
+      "wer": 0.6416666666666667
+    },
+    {
+      "subset_type": "accent",
+      "subset": "bermuda",
+      "cer": 0.2449438202247191,
+      "wer": 0.5476190476190477
+    },
+    {
+      "subset_type": "accent",
+      "subset": "ireland",
+      "cer": 0.1690856313497823,
+      "wer": 0.45569620253164556
+    },
+    {
+      "subset_type": "accent",
+      "subset": "newzealand",
+      "cer": 0.14555256064690028,
+      "wer": 0.375
+    },
+    {
+      "subset_type": "accent",
+      "subset": "other",
+      "cer": 0.24492979719188768,
+      "wer": 0.5405405405405406
+    },
+    {
+      "subset_type": "accent",
+      "subset": "us",
+      "cer": 0.17611423626135872,
+      "wer": 0.44368026898976004
+    },
+    {
+      "subset_type": "accent",
+      "subset": "african",
+      "cer": 0.22567703109327983,
+      "wer": 0.5144508670520231
+    },
+    {
+      "subset_type": "accent",
+      "subset": "scotland",
+      "cer": 0.20196353436185133,
+      "wer": 0.475177304964539
+    },
+    {
+      "subset_type": "accent",
+      "subset": "singapore",
+      "cer": 0.2192513368983957,
+      "wer": 0.5454545454545454
+    },
+    {
+      "subset_type": "accent",
+      "subset": "canada",
+      "cer": 0.211519364448858,
+      "wer": 0.4887640449438202
+    }
+  ]
+}

cv7_transcriptions.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5a66ac5d7066deea765b173c0c20e99a486242b4b406f1badb45cab3a50f7e05
+size 121108847

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 sha256:73fef16d13ebaed5b7bf658d0496096f26536dcdad4cdb352f18ed289f934bdb
+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/tmpvcna4y2j", "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}